fix(tui): harden all Box TUI tabs against nulls, boundary slices, and small screens

- netvm-chrome.sh --contained: bwrap fs jail inside netns (profile
  home only + vault ro at /tmp/vault, Chromium sandbox stays on)
- netvm-proton.sh: run proton-cli as the profile identity in netns
  + jail (config dir + static binary, PROTON_NO_INPUT=1)

.agents/skills/box/SKILL.md

@@ -1,50 +0,0 @@
----
-name: box
-description: Use the box CLI to check NetVM fleet health and read the latest from each agent.
----
-
-# Box Fleet CLI
-
-Use `box` (`/usr/local/bin/box`, the NetVM unified orchestrator CLI) for all fleet observation and agent coordination. Prefer read-only lookups first; coordinate via sidechats, never Main Chat dumps.
-
-Nodes (node == agent == profile): `muse`, `pip`, `646`, `opm`, `def`, `dev`.
-
-## Latest From Each Agent (Default Workflow)
-
-1. `box fleet status` — node health, CDP status, active page/thread.
-2. `box lookup unread` — unread counts across agents.
-3. `box lookup threads` — registered threads and sidechats.
-4. Per agent with activity: `box thread list <agent>`, then `box thread view <agent> <thread_id> --limit 10` (use the thread UUID from the list; `main` only for urgent human-visible items).
-5. `box dm log -n 20` (or `--agent <agent>`) — recent inter-agent DMs, work orders, and acks.
-6. `box approvals check` — agents blocked on browser or key approval.
-
-Add `--json` to any command for machine-readable output when parsing results in scripts.
-
-## Common Commands
-
-- `box fleet status` / `box fleet cdp <node>` — health table / CDP endpoint plus SSH forward.
-- `box fleet heal <node>` — diagnose + fix + verify a node (lock, watchdog timer, Warp tunnel, egress, relay).
-- `box watchdog status` / `box watchdog run <node|relay>` — timer states + evidence / trigger an immediate watchdog run.
-- `box thread list [<agent>]` — threads for one agent, or all fleet sidechats when omitted.
-- `box thread view <agent> <thread_id|main> --limit N` — recent messages from one thread.
-- `box dm log -n N [--agent X] [--filter TEXT]` — recent DM activity.
-- `box dm send --agent <self> --to <peer> --target <sidechat> "<msg>"` — peer DM.
-- `box lookup summary|fleet|threads|unread|approvals` — seamless one-shot lookups.
-- `box job list` / `box job log` — scheduled jobs and execution events.
-- `box harvest status` / `box followup list` — harvest watermarks / pending nudges.
-- `box muse-choices on|off|status|logs|reconcile|resolve` — Muse TUI auto-answer daemon switch, state, per-pane logs, held-prompt resolve (default on; `off` is the box-command opt-out).
-- `box runtime list|send|launch|layout|spread` — Muse CLI tmux runtimes: live state + approval posture, send-keys input, auto-approved launches, pane-geometry layout + spread for squeezed panes.
-- `box tmux tally` / `box tmux auto [status|on|off|watch|once|logs|match]` — multi-socket Tmux worker tally, regex auto-approver daemon & guardrails.
-- `box onboard connects` / `box onboard-tui` — fleet & client onboarding inventory, CDP ports, OTP salvage & 4-surface TUI.
-- `box invite status|code <node>|redeem <node> <CODE>` / `box usage [--node N]` — invite codes and usage limits.
-- `box kpi status|report <node>|routes|spawn-worker|auto-spawn` — fleet KPI tracking, spend/limit metrics, route health, runtime preservation advisories, and background worker auto-spawning.
-- `box chromebox permissions <node> list|get <t>|set <t> <v>|describe <tab>` — settings-menu toggles (readback-verified sets).
-
-## Rules
-
-- Sidechat-first per `CHAT_POLICY.md`: `646 tasks`, `heartbeat` (opm), `646-pip-coord`, `646-opm-coord`. Never route routine checks or coordination through `main`.
-- Never paste multi-KB logs or dumps into chat; write payloads under `logs/` and send a short path pointer instead.
-- Avoid blocking commands in automated runs: `box fleet watch`, `box dm tail`, `box approvals watch`, `box dm chat` (interactive REPL).
-- `box` probes CDP per node and can take several seconds; use generous timeouts and `--json` for scripted use.
-- Preserve runtime and quota limits: prefer spawning subagents (`jobs/`, `subagent_tracker`) or background workers (`box kpi spawn-worker`) over long conversational prose to avoid `VANITY_IDLE` flags.
-

.gitignore

@@ -1,30 +0,0 @@
-# NetVM Transfers & Local State
-transfers/
-*.log
-logs/
-*.jsonl
-pipelines.json
-followups.json
-siphon-watermarks.json
-identity-state.json
-review/
-
-__pycache__/
-*.pyc
-*.bak
-*.bak-*
-*.orig
-keepalive-config.json
-main-chat-watchdog.state
-strategy.json
-variables.json
-chained-jobs.json
-.state/
-*.lock
-*watermark*
-ssl/
-job-scheduler-state.json
-var/
-swarms.json
-subagent-sessions.json
-conversation-nudge-tracker.json

ACCOUNTS.md

@@ -1,67 +1,51 @@
-# NetVM Account Registry
+# Login registry — secret-free
 
-One row per agent. All data in columns — no joins, no translation.
-The `agent` name is the canonical identifier used everywhere:
-node name, chrome-box profile, API `--account`, and the agent's display name.
+Which product login lives in which chrome-box profile, on which NetVM node,
+with which egress, in what auth state. This is structure only: **no
+passwords, no tokens, no session cookies, no OTP codes — ever.** Credential
+pointers at most (e.g. "human", "credential-gateway:<id>").
 
-## Schema
+The 1:1 chain: `login -> profile = node = Warp identity = veth/CDP slot =
+consistent egress`. Network details live in NODES.md; this file maps the
+human side (whose login, what for, does it work).
 
-| Column | Description |
-|--------|-------------|
-| agent | Canonical name. Used for node, profile, API account. |
-| node | NetVM node name (== agent). |
-| profile | Chrome-box profile (== agent). |
-| login_type | How this session was authenticated: `email_otp`, `phone_otp`, `password`, `instagram` |
-| meta_label | Label shown in Meta account selector (e.g., "Meta Account", "piparada") |
-| email | Email used for login (if email_otp). Never store passwords. |
-| phone_otp | `yes` if phone OTP was used. Never store the phone number. |
-| instagram_linked | `yes`/`no`/`unknown` — whether Meta account has IG linked |
-| status | `active`, `pending_auth`, `needs_signup`, `expired`, `disabled` |
-| egress_ip | Current WARP egress IP for the node |
-| cdp_port | CDP port for the browser |
-| display_name | Agent's chosen display name in muse.ai (may differ from `agent`) |
-| notes | Freeform context |
+## Auth states
 
-## Accounts
+| state | meaning | who moves it |
+|-------|---------|--------------|
+| `pending-identity` | profile exists, no Warp identity yet | human runs `netvm-new-identity.sh <profile>` |
+| `pending-auth` | node up, nobody logged in yet | human logs in (browser or credential gateway) |
+| `2fa-pending` | login needs a human 2FA/OTP step | human via ethical-captcha handoff; OTP routed by email-alert |
+| `active` | logged in, session healthy | operator verifies; automation may proceed |
+| `expired` | session died | back to `pending-auth` (human) |
+| `retired` | login no longer used | operator tears down node, archives row |
 
-| agent | node | profile | login_type | meta_label | email | phone_otp | instagram_linked | status | egress_ip | cdp_port | display_name | notes |
-|-------|------|---------|------------|------------|-------|-----------|------------------|--------|-----------|----------|--------------|-------|
-| muse | muse | muse | email_otp | ltd.pixels.ltd@gmail.com | ltd.pixels.ltd@gmail.com | no | unknown | active | 104.28.195.181 | 9410 | muse | Main dev agent. Logged in 2026-10-03 via email OTP on bl. |
-| pip | pip | pip | phone_otp | piparada | io.antonio.parada@gmail.com | yes | yes | active | 104.28.195.181 | 9420 | pip | Phone OTP login. Linked with IG piparada, email io.antonio.parada@gmail.com. Verified active 2026-10-04. |
-| 646 | 646 | 646 | phone_otp | Meta Account | - | yes | no | active | 104.28.195.181 | 9430 | 646 | Shares phone number with piparada's account. Logged in 2026-10-03 via phone OTP (first Meta Account option). Node created 2026-10-03 (warp-646, CDP 9430). Browser up, session active. |
-| def | def | def | email_otp | defnotabotnet@gmail.com | defnotabotnet@gmail.com | no | yes | active | 104.28.195.181 | 9450 | def | Full onboarding completed 2026-10-04; age verification cleared via Instagram linking (paradahub). Active chat session. |
-| opm | opm | opm | email_otp | Nico Parada | artglobal.cc@gmail.com | no | yes | active | 104.28.195.181 | 9440 | opm | Email changed from yourfriendnico@proton.me to artglobal.cc@gmail.com. Linked with IG auxfate. Browser up, session active. |
-| dev | dev | dev | email_otp | paradaproduced@gmail.com | paradaproduced@gmail.com | no | yes | active | 104.28.195.181 | 9460 | dev | Full onboarding completed 2026-10-04; unlocked /access gate via Meta Accounts Center IG linking (veryraremeta). Active chat session. |
+Operators never create or touch credentials. If it creates or touches a
+credential, it is human-only. Everything else, operators handle.
 
-## Login Type Details
+## Registry
 
-### email_otp
-- Flow: Enter email → Receive OTP via email → Enter OTP → Select account (if multiple)
-- Used by: muse
-- Credentials: Email address (stored). OTP is transient.
+| login | product | profile/node | purpose / owner | auth state | 2FA / verify route | notes |
+|-------|---------|--------------|-----------------|------------|--------------------|-------|
+| — | — | tp | orchestrator / operator-main | pending-auth | — | first node; no product login yet |
+| — | — | smoke | muse-646-patha | active | — | 646's profile; muse.ai login completed 2026-10-03 |
 
-### phone_otp
-- Flow: Enter phone → Receive SMS OTP → Enter OTP → Select Meta account (if multiple)
-- Used by: pip, 646
-- Credentials: Phone number is NEVER stored (PII). Only `phone_otp=yes` flag.
-- Note: One phone number can map to multiple Meta accounts (observed: 2 accounts).
+## Known login flows
 
-### Meta Account Selection
-When a phone number maps to multiple Meta accounts, muse.ai shows a selector:
-- Screenshot: `docs/meta-account-selection.png`
-- Each option is a SEPARATE Muse container (not linked profiles).
-- The `meta_label` column records which option was selected.
-- Instagram-linked accounts show IG avatar in selector.
+### muse.ai (recon 2026-10-03, via CDP DOM)
+- Homepage has "Log in" buttons (JS, no href). Click -> inline form, same URL.
+- "Log in or create an account" — single field: "Mobile number or email (required)" + Continue.
+- Phone/email OTP flow (SMS or email code). No password, no OAuth buttons.
+- Human completes it in one visible session; operators verify + automate after.
 
-## Naming Convention
+## Provisioning a new login (dev)
 
-**Rule:** The `agent` column value is used identically for:
-- NetVM node name (`/etc/netvm/<agent>.conf`)
-- Chrome-box profile (`~/.local/share/chrome-box/profiles/<agent>/`)
-- API account (`muse-chat-api.py --account <agent>`)
-- CDP port mapping (deterministic per agent)
-
-**Exception:** `display_name` may differ (user-chosen in muse.ai UI).
-Example: agent `pip` has display_name `pip` (renamed from 'Muse').
-
-Do NOT use different names for node vs profile vs API. That causes bugs.
+1. Operator: `chrome-box create <profile>` (profile name = future node name).
+2. Human: `netvm-new-identity.sh <profile>` (Warp identity — credential).
+3. Operator: `netvm-node-up.sh <profile>`; add rows to NODES.md and here
+   (`pending-identity` -> `pending-auth`).
+4. Human: authenticate the login in the profile's browser
+   (`netvm-chrome.sh <profile>` visible, or credential-gateway injection).
+   Row -> `active`.
+5. Operator: verify with `netvm-exec.sh <profile> -- ...` / CDP; keep the
+   session warm. On 2FA: ethical-captcha handoff, OTP via email-alert.

CHAT_POLICY.md

@@ -1,76 +0,0 @@
-# NetVM Agent Chat Policy: Main Chat Preservation
-
-**Status:** ACTIVE POLICY (Mandatory across all fleet automation, jobs, and operator tooling)  
-**Date:** 2026-10-04  
-**Version:** 2.0  
-**Applies to:** All autonomous agents (`muse`, `pip`, `646`, `opm`), scheduled jobs (`super job`), orchestrator tooling (`super dm`, `dm.py`), and human operators.
-
----
-
-## 1. The Core Principle: Main Chat Is Sacred
-
-> **Rule:** **Avoid using Main Chat whenever possible.**
->
-> When Main Chat gets bogged down with automated entries, scheduled job triggers, log dumps, or inter-agent chatter, **work stops actually getting done**.
-> Browser DOM virtualizers lag or crash, context windows saturate with noisy outputs, and the agent's attention drifts away from primary operator directives.
-
-Main Chat is reserved **exclusively** for high-level human operator oversight, urgent human-visible escalations, and direct operator conversational alignment.
-
----
-
-## 2. Channel Segregation Rules
-
-### Rule A: Scheduled Jobs MUST Target Dedicated Sidechats
-* **Never** configure a routine scheduled job (e.g., cron checks, health monitors, heartbeats, periodic scrapes) to deliver to `main`.
-* Every job definition in `jobs/<name>.json` must explicitly specify:
-  * `"dm_target": "<sidechat-name-or-uuid>"` OR
-  * `"sidechat": { "create": true, "name_template": "...", "reuse_key": "..." }`
-* Any job found dumping routine health outputs or telemetry into `main` must be immediately migrated to a dedicated sidechat.
-
-### Rule B: Inter-Agent Communication Runs via Sidechats / Side Agents
-* Autonomous agents communicating with one another (e.g., `646` ↔ `pip`, `opm` ↔ `646`) must use dedicated coordination sidechats (e.g. `646-pip-coord`, `646-opm-coord`, `646 tasks`).
-* Do not route peer coordination or sub-task requests through an agent's Main Chat.
-* Subordinate or delegated tasks should be spun off to side agents or sidechats to isolate the conversation state and prevent main thread contamination.
-
-### Rule C: Large Payloads Transferred via File System, Not Chat
-* Do **not** dump multi-kilobyte log extracts, raw HTTP responses, table dumps, or diffs into any chat window.
-* Payloads must be written to disk on `bl` or the VM (e.g. in `/home/super/Projects/NetVM/logs/` or `/srv/box/`) and referenced via short path / pointer in the message:
-  * ✅ *Good:* `[RESULT 12345] Health check completed. 6/6 endpoints OK. Detailed breakdown saved to logs/http-health-20261004.log`
-  * ❌ *Forbidden:* Pasting 200 lines of raw curl outputs or JSON logs into chat.
-
-### Rule D: Operator CLI (`super dm`) Enforces Sidechat-First Flow
-* Interactive conversational sessions (`super dm chat <agent>`) prompt for or default to sidechats and issue an explicit policy warning whenever Main Chat is selected.
-* When dispatching one-off DMs via `super dm send` or `super dm wo`, operators must prefer `--target "<sidechat>"` over `--target main`.
-
----
-
-## 3. Fleet Addressing Directory
-
-| Target | Agent | Purpose | Policy Tier |
-|---|---|---|---|
-| `main` | All (`muse`, `pip`, `646`, `opm`) | Direct human-to-operator urgent interventions only | **Restricted / Minimal** |
-| `heartbeat` | `opm` | Automated DM pipeline heartbeat verification | **Sidechat Required** |
-| `646 tasks` | `646` | Daily check-ins, execution health, operator tasks | **Sidechat Required** |
-| `646-pip-coord` | `pip` / `646` | Peer coordination between pip and 646 | **Sidechat Required** |
-| `646-opm-coord` | `opm` / `646` | Peer coordination between opm and 646 | **Sidechat Required** |
-
----
-
-## 4. Violations & Enforcement
-
-1. **Dispatcher Guard:** Scheduled jobs with `schedule != "manual"` and no `dm_target` or `sidechat` configuration will be audited and retrofitted with dedicated sidechat targets.
-2. **Review Checklist:** Any PR, skill, rule, or script introducing automated messages must verify that output lands in a sidechat or log file, never in Main Chat.
-
----
-
-## 5. Changelog
-
-### v2.0 — 2026-10-04: Sidechat-only enforcement
-* **New rule:** No DM lands in Main Chat unless explicitly authorized. `dm.py send` / `job-dispatch.py` refuse `--target main` without `--allow-main-chat` (exit 2, `main_chat_blocked` log event, before any browser navigation). Job JSON opt-in key: `"allow_main_chat": true`.
-* **Watcher:** `main-chat-watchdog.py` (every 5 min) classifies dm-log events into blocked/ authorized / violation.
-* **Triage:** `docs/SIDECHAT-POLICY-TRIAGE.md` `— where to look first on failure.
-* **Fix:** `SIDCHAT_ALIASES["heartbeat"]` de-collided — was pointing at 646's tasks thread (`1e75a740-...`); now points at the dedicated heartbeat sidechat (`0077e918-...`, reuse_key `heartbeat-opm` in `job-sidechats.json`).
-
-### v1.0 — 2026-10-04: Initial policy
-* Main Chat preservation: scheduled jobs and inter-agent communication must use dedicated sidechats.
-* Fleet addressing directory established.

CLIENT-ONBOARDING-RUNBOOK.md

@@ -1,146 +0,0 @@
-# Client Onboarding & Fleet Runbook (`CLIENT-ONBOARDING-RUNBOOK.md`)
-
-## 1. Overview & Agency Context
-This document defines the complete standard operating procedure (SOP) and automated runbook for provisioning, authenticating, and onboarding client agency profiles (nodes) into the NetVM multi-tenant fleet on `bl`.
-
-In accordance with the NetVM Ethics Charter (`https://start.muse-dev.online/ethics.html`):
-- Managed services are strictly run for consenting clients with explicit authority.
-- Every client receives a completely isolated network namespace (`warp-<node>`), dedicated WireGuard tunnel identity, isolated Chrome profile, and separate credentials.
-- Canonical Naming Convention: `node == agent == profile == API account`.
-
----
-
-## 2. Fleet Architecture & Port Allocation
-
-The fleet uses a deterministic `94x0` CDP port and `warp-<node>` naming convention:
-
-| Node | CDP Port | Netns | Egress IP | Purpose / Profile |
-|------|----------|-------|-----------|-------------------|
-| `muse` | `9410` | `warp-muse` | Dedicated WARP | Primary Dev / Orchestrator |
-| `pip` | `9420` | `warp-pip` | Dedicated WARP | Production Agent |
-| `646` | `9430` | `warp-646` | Dedicated WARP | Production Agent |
-| `opm` | `9440` | `warp-opm` | Dedicated WARP | Production Agent |
-| `def` | `9450` | `warp-def` | Dedicated WARP | Production Agent |
-| `<new>` | `9460+` | `warp-<new>`| Dedicated WARP | Next provisioned client node |
-
----
-
-## 3. Step-by-Step Client Onboarding SOP
-
-### Phase 1: Infrastructure Provisioning (Automated)
-Run the idempotent node provisioning script to generate the WireGuard identity, network namespace, CDP relay, and chrome-box profile:
-
-```bash
-# Example: Provisioning node 'dev1'
-./bin/netvm-provision-node.sh dev1
-```
-*Verification:*
-- Namespace created: `ip netns list | grep warp-dev1`
-- Registry updated in `NODES.md` and `ACCOUNTS.md`.
-
----
-
-### Phase 2: Sign-in Initiation (`super cred initiate`)
-Launch the client login flow without handling raw passwords or secrets:
-
-```bash
-# For email OTP login:
-super cred initiate --node dev1 --email client@domain.com
-
-# Or via Python Agent API:
-python3 bin/cred-client.py initiate --node dev1 --email client@domain.com
-```
-
-- If already authenticated, exits `0` (`active`).
-- If awaiting verification code, exits `2` (`awaiting_otp`).
-
----
-
-### Phase 3: Submitting Transient OTP (`super cred submit-otp`)
-When the client or operator receives the 6-digit email OTP:
-
-```bash
-super cred submit-otp --node dev1 --otp 123456
-```
-
-- The code is submitted transiently and is never persisted to disk or logs.
-- If the account directly enters chat, status transitions to `active`.
-- If the account requires age verification, it advances to Phase 4.
-
----
-
-### Phase 4: Resolving the Age Verification Gate (`/access/verification`)
-When a brand-new or unlinked client profile reaches the Muse age verification gate:
-
-#### Method A: Instagram Linking (Recommended)
-1. Run:
-   ```bash
-   super cred link-instagram --node dev1 [--notify]
-   ```
-2. The system provides a one-tap Tailscale portal URL:
-   `http://bl.tailfb5960.ts.net:8765/verify/dev1`
-3. **Crucial Rule**: The operator or client must link an **established / aged Instagram profile** (not created within minutes). Brand-new Instagram accounts lack mature age signals, causing Meta Accounts Center to disable the Confirm button.
-4. If completed via mobile/desktop browser, use an Incognito/Private window to prevent ambient Meta cookie bleed.
-5. If executing automated RPA in-browser, inject the Instagram credentials and security code directly into the container's CDP session.
-
-#### Method B: Credit Card Verification (Fallback)
-If Instagram linking is not available, operator can complete the verification using a client payment card on `/access/verification`.
-
----
-
-### Phase 4.1: Edge Gate — Hard Audience Lockout (`/access` vs `/access/verification`)
-- **Observed Behavior**: If an account routes to `https://muse.ai/access` with the text `"Muse isn't available to all audiences."` instead of `https://muse.ai/access/verification`:
-  - The Meta account is temporarily unverified or lacks linked identity signals.
-  - The in-app endpoint (`/api/hatch/age-confirmation/linking-web-auth`) returns `403 Forbidden`.
-  - Reloading or navigating directly to `/` or `/access/verification` immediately redirects back to `/access`.
-- **Root Cause**:
-  - Meta accounts without an active linked profile (Facebook or Instagram) trigger Meta's general audience filter on Muse before the conversational AI product can be initialized.
-  - In addition, attempting automated sign-in on low-reputation / unverified identities directly from server/VPN IPs will trigger Google reCAPTCHA Enterprise checkpoints (`auth_platform/recaptcha`).
-- **Proven Unblocking SOP (The Direct Meta Accounts Center Flow)**:
-  1. Open a clean browser session with the target Meta Account signed in (`https://accountscenter.meta.com/`).
-  2. Navigate to **Accounts** → **Add Accounts** (`/add_accounts/`).
-  3. Enter the Instagram credentials for an older/established IG profile (`veryraremeta`, `paradahub`, etc.) and submit any required 2FA/email OTP.
-  4. If returned to Accounts Center, click **Add Instagram** again to initiate the OAuth handoff:
-     `https://www.instagram.com/fxcal/auth/login/?app_id=633385687760560...&flow=igcalcomet&entry_point=frl_web_settings`
-  5. On the *"Meta needs to access info from your Instagram account"* prompt, click **[Continue]**.
-  6. Accounts Center returns to the confirmation screen (`/add/?auth_flow=ig_linking&token=...&blob=...`) → click **[Confirm]**.
-  7. Meta sends security confirmation: *"Did you just move your profiles into the same Meta Account?"*.
-  8. Once confirmed in Accounts Center, simply navigate back or reload `https://muse.ai/` inside the NetVM node. The `/access` lockout drops immediately, and the node enters active chat (*"Hey! I'm your personal agent..."*).
-
-### Phase 4.2: Automated CDP Meta Linking Bot
-When an edge-gate is detected or when linking a fresh client profile:
-1. Trigger the automated CDP driver:
-   ```bash
-   sudo ip netns exec warp-<node> python3 /home/super/Projects/NetVM/bin/meta-acct.py link-instagram <node>
-   ```
-2. The bot:
-   - Navigates headless Chromium to `https://accountscenter.meta.com/manage/`.
-   - Locates and clicks **Add profiles and devices**.
-   - Selects the Instagram cross-app linking flow.
-   - Automatically navigates to the `frl_web_settings` FXCAL OAuth grant.
-3. Once completed or after submitting Instagram credentials, re-query the account state:
-   ```bash
-   super cred meta-audit --node <node>
-   ```
-
----
-
-### Phase 5: Vitality & Status Monitoring
-Query individual or fleet-wide health:
-
-```bash
-# Check single node
-super cred status --node dev1
-
-# Check entire fleet
-super cred list
-```
-
----
-
-## 4. Rate-Limiting & Operational Safety Rules
-
-To avoid platform anti-automation challenges and maintain high reputation:
-1. **Pacing / Spacing**: Space new node creations and Instagram authorizations by **at least 15–20 minutes** per IP/session.
-2. **Namespace Isolation**: Never attempt multi-account auth inside the same browser profile. Always execute inside the client's dedicated `warp-<node>` netns.
-3. **No Credential Logging**: Never print plain text passwords or authentication tokens to stdout, git-tracked markdown, or plain text logs.

CREDSTORE.md

@@ -1,92 +0,0 @@
-# Meta Credential Store (operator-only)
-Centralized encrypted store for Muse, Instagram, Facebook account credentials.
-## Location (VM only)
-- `/etc/netvm/meta-credentials/store.age` — age-encrypted JSON (600 root)
-- `/etc/netvm/meta-credentials/.age-key` — age private key (600 root)
-- `/usr/local/bin/meta-creds.sh` — CLI (700 root)
-## Usage
-```bash
-sudo meta-creds.sh list muse              # list account IDs (no secrets)
-sudo meta-creds.sh get muse <id>          # output JSON (never log this)
-sudo meta-creds.sh add muse <id>          # interactive prompts
-```
-## Naming
-
-Store ID == `ACCOUNTS.md` `agent` name (e.g. `646`, `pip`, `muse`). The
-secret store and the secret-free registry join on this ID — same account,
-different jobs (secrets vs. state).
-
-## Schema
-```json
-{
-  "muse": {
-    "<id>": {
-      "email": "...",
-      "phone": "...",
-      "via_meta_account": "<facebook|instagram id>",
-      "age_verified": "true",
-      "instagram_linked": "<handle>",
-      "verified_by": "human", "verified_at": "2026-10-03T...",
-      "notes": "..."
-    }
-  },
-  "instagram": {
-    "<id>": {
-      "username": "...", "password": "...",
-      "email": "...", "phone": "...",
-      "accounts_center": "<alias>",
-      "linked_to": ["<other store id>", "..."],
-      "login_methods": ["password", "phone_otp"]
-    }
-  },
-  "facebook": {
-    "<id>": {
-      "email": "...", "password": "...", "phone": "...",
-      "accounts_center": "<alias>",
-      "linked_to": ["<other store id>", "..."],
-      "login_methods": ["password", "phone_otp"]
-    }
-  }
-}
-```
-
-### Field notes
-
-- `phone`: mobile number for login/2FA. **May be stored here** (encrypted);
-  one phone can map to multiple accounts (observed: 646 + piparada share
-  a number) — never treat it as a unique key.
-- `via_meta_account` (muse): when muse.ai auth runs through a Meta
-  account (phone OTP → Meta account → muse.ai), points at the
-  `facebook`/`instagram` entry. This is the 646/pip intersection.
-- `accounts_center`: local alias for the Accounts Center (e.g. `ac-646`).
-  Post early-2026 this is the login blast-radius boundary — every account
-  in one Center logs into every other by default.
-- `linked_to`: other store IDs in the same Accounts Center. Cached from
-  the Meta API's `list-linked`; **the API is ground truth** — when they
-  disagree, the API wins and the store gets updated.
-- `login_methods`: how the account can be authenticated. Drives which
-  flow the automation attempts.
-
-## Intersections
-
-- **Phone ↔ accounts (1:many):** the store holds the number (encrypted);
-  the human is no longer the sole holder, but the number still never
-  appears in logs, chat, memory, or the secret-free registry.
-- **Meta credential → muse.ai session:** a `facebook`/`instagram` entry
-  can be the auth path for a `muse` login. Follow `via_meta_account`.
-- **Store ↔ ACCOUNTS.md:** joined on ID. Store = secrets, registry = state.
-- **Store ↔ Meta Accounts Center API** (`docs/META-ACCOUNTS-API.md`):
-  the API reads linkage ground truth; the store caches it in `linked_to` /
-  `accounts_center`.
-
-## Rules
-
-- Operators only. Developers never get access (prevents board leaks).
-- Decrypt transiently, never log values, never put in chat/memory.
-- Phone numbers and PII **may** live in `store.age` (age-encrypted, 600
-  root, VM only). They must never appear in plaintext anywhere else:
-  no logs, no chat, no memory, no registry, no board.
-- Human validates Instagram linking and Meta account ownership;
-  operators automate after.
-- When adding an account, fill `accounts_center` / `linked_to` from the
-  Meta API (`list-linked`), not from memory.

INFRA.md

@@ -1,50 +0,0 @@
-# NetVM Infrastructure Layout
-
-## bl (100.123.153.75) — Main Compute
-- 16 cores, 28GB RAM
-- Roles: Browser automation, NetVM nodes
-- Access: VM -> bl via SSH
-- WARP: Per-node identities in /etc/netvm/
-- Browsers: Headless Chromium via chrome-box
-- API: muse-chat-api.py via netvm-exec (CDP)
-- Sign-in: muse-signin.py --email <addr> [--otp <code>]
-- Hygiene: netvm-reaper.sh
-- Approvals: In-browser via chat (see below)
-
-## VM (34.139.37.135) — Gateway + Vault
-- E2 micro (2 vCPU, 1GB RAM)
-- Roles: Jump host to bl, credential store
-- Credential store: /etc/netvm/meta-credentials/ (age-encrypted)
-- No browser automation (resource constraints)
-
-## Laptop — Dev
-- Roles: Iteration, visible browser debugging
-- Nothing production
-
-## Credential Flow
-- Meta accounts: /etc/netvm/meta-credentials/store.age (VM)
-- WARP identities: /etc/netvm/node.conf (per-machine, root 600)
-- Operators handle transiently; never log values.
-
-## In-Browser Approvals
-When automation needs human input (OTP, confirmation):
-1. Script exits with code 2 and prints "APPROVAL_NEEDED: <details>"
-2. Operator sees this and asks user via chat
-3. User provides input (e.g., OTP code)
-4. Operator re-runs script with --otp <code>
-5. Script completes the flow
-
-No file-based queue needed — the chat IS the approval interface.
-The human is already in the chat; the automation just needs to
-signal when it's stuck.
-
-## Sign-In Flow (muse-signin.py)
-Automated login for muse.ai accounts:
-- Step 1: Check if already logged in (skip if yes)
-- Step 2: Click "Log in"
-- Step 3: Enter email
-- Step 4: Click "Continue"
-- Step 5: Detect OTP prompt
-  - If --otp provided: enter it, click Next, verify
-  - If not: exit 2 with APPROVAL_NEEDED
-- Credentials never stored; OTP is transient.

INSTAGRAM-CRED-POOL.md

@@ -1,126 +0,0 @@
-# Instagram Credential Pool Specification (`INSTAGRAM-CRED-POOL.md`)
-
-## 1. Overview & Agency Context
-When onboarding agency client profiles ("nodes") to Muse (`muse.ai`), new accounts and certain unconfirmed email accounts encounter the post-OTP Age Verification gate (`/access/verification`).
-
-While credit-card verification is not suitable for autonomous multi-tenant operations, **Meta OAuth / Instagram Linking** is the highest-reliability verification route. 
-
-Currently, the system uses a **Human-in-the-Loop** model:
-- An operator receives an authorization notification via Tailscale / email.
-- The operator signs into Instagram via a one-tap link from their mobile device or laptop.
-
-This specification details the future architecture for **Automated Instagram Credential Pooling** to remove human intervention entirely while strictly complying with the NetVM Ethics Charter (`https://start.muse-dev.online/ethics.html`).
-
----
-
-## 2. Architecture & Design Principles
-
-### 2.1 Isolation & Multi-Tenancy (Ethics Charter Compliant)
-- **1-to-1 Node Mapping**: Each client node (`node == agent == profile == API account`) maintains dedicated browser state, WireGuard netns isolation (`warp-<node>`), and separate credentials.
-- **Dedicated IG Identities**: Instagram accounts in the pool are provisioned specifically for age-verification linking, never shared concurrently across different active client profiles.
-- **Encrypted Secret Storage**: Instagram credentials (username, password, 2FA TOTP secret, session cookies) are stored in an encrypted credential vault (e.g., `age`-encrypted `/etc/netvm/meta-credentials/store.age`), never committed in plain text to git or unencrypted markdown.
-
-### 2.2 Pool States & Lifecycle
-
-```mermaid
-stateDiagram-v2
-    [*] --> Available : Provisioned & Verified
-    Available --> Assigned : Reserved for Node Onboarding
-    Assigned --> Linking : Navigating Meta OAuth in netns
-    Linking --> Linked : Age Gate Cleared on Muse
-    Linking --> Cooloff : Checkpoint / Rate-limit Hit
-    Cooloff --> Available : Cooldown Elapsed
-    Linked --> InUse : Node Active in Fleet
-```
-
-- **`available`**: Account is verified, healthy, and not currently tied to any active Muse profile.
-- **`assigned`**: Temporarily reserved by `super cred` for onboarding node `<node>`.
-- **`linking`**: Automated driver navigating the Meta Accounts Center flow inside the isolated namespace.
-- **`linked`**: Successfully bound to Muse account.
-- **`cooloff`**: Encountered challenge or cooldown; resting before re-qualification.
-
-### 2.3 Meta Account Center Constraints & Edge Cases
-- **1-to-1 Linking Constraint**: Meta Accounts Center rejects linking if the target Instagram account is already associated with an existing Meta / Muse profile (`auth_flow=ig_linking` drops to `add_accounts` error page with `token` and `blob` parameters).
-- **Brand New Account Age Gate Limitation**:
-  - Newly created Instagram accounts without a mature age/identity verification tier or age signal trigger Meta Accounts Center to disable the **"Confirm"** action (`aria-disabled="true"` on `/add_accounts/?flow=HATCH_AGE_VERIFICATION_IG_UPSELL`).
-  - Meta Accounts Center uses Instagram accounts for age verification by checking that the linked Instagram profile itself has established age signals. A freshly minted account created minutes prior lacks this profile history, leaving the age verification unsatisfied.
-  - **Agency Recommendation**: Pre-aged or verified Instagram identities in the pool with established age badges/profiles, or using established client identities, rather than accounts created in the immediate transaction.
-- **Dormant / "Ghost" Account Lockout Mode (`/access` Hard Exclusion & Resolution)**:
-  - An account that has chronological calendar age (e.g. created ~5 months ago) but has **zero posts, zero regular engagement, and no established social graph** can fail Meta's automated audience eligibility check completely upon Muse onboarding, routing to `https://muse.ai/access` (*"Muse isn't available to all audiences"*).
-  - Attempting automated sign-in on these low-reputation identities from datacenter/VPN egress IPs triggers Google reCAPTCHA Enterprise checkpoints.
-  - **The "Add Again" Two-Step Handoff Resolution**:
-    1. Sign in to `https://accountscenter.meta.com/` using the cached Meta Account session.
-    2. Add Account -> complete Instagram sign-on & OTP.
-    3. Returning to Meta Accounts Center, click **Add Instagram a second time** (`ADD AGAIN`).
-    4. Meta generates the OAuth handoff URL:
-       `https://www.instagram.com/fxcal/auth/login/?app_id=633385687760560&etoken=...&next=https%3A%2F%2Faccountscenter.meta.com%2Fadd%2F%3Fauth_flow%3Dig_linking%26background_page%3D%252Fmanage&flow=igcalcomet&entry_point=frl_web_settings&initiator_fbid=...`
-    5. Prompt displays: `"[<instagram_handle>] Meta needs to access info from your Instagram account. [Continue] [Not You?]"`.
-    6. Clicking **[Continue]** redirects to Accounts Center with query parameters `token` and `blob` (`/add/?auth_flow=ig_linking&token=...&blob=...`).
-    7. Clicking **[Confirm]** completes the account merge, prompts Meta's confirmation email (*"Did you just move your profiles into the same Meta Account?"*), and **instantly clears the `/access` block on Muse**, transitioning the session into active chat.
-- **Session Bleed & OIDC Secondary Auth Trip (`auth.meta.com`)**:
-  - When the link is opened in a browser that has existing Meta session cookies (e.g. from Facebook, Oculus, or another Meta account), selecting the new Instagram identity triggers a secondary OpenID Connect reconciliation trip (`https://auth.meta.com/?waterfall_id=...&redirect_uri=auth.meta.com/oidc/...&source_app_id=633385687760560`).
-  - This prompts the user with **"Log in with your Meta account"** because the browser's ambient Meta session does not match the freshly authenticated Instagram identity.
-  - If the user confirms with their cached personal Meta credentials, Meta attempts to merge/link across two disparate account graphs, creating an authorization loop or conflict.
-  - **Resolution**: The link must strictly be opened in an **Incognito / Private window** or a completely clean browser profile with zero cached Meta/Facebook/Instagram cookies.
-- **In-Namespace Isolation**: Automated pool linking runs in Chromium directly inside `warp-<node>` with an isolated profile, avoiding cross-session cookie collisions entirely.
-
----
-
-## 3. Automated Driver Mechanics
-
-### 3.1 Fetching Authorization Payload
-From the node's running browser tab sitting on `/access/verification`:
-```javascript
-const res = await fetch('/api/hatch/age-confirmation/linking-web-auth?account_type=instagram', {
-  headers: { 'Accept': 'application/json' }
-}).then(r => r.json());
-// res.url: https://www.instagram.com/fxcal/auth/login/?app_id=...&next=...
-```
-
-### 3.2 Automated Headless Linking Flow
-1. Rather than opening a blocked popup, the driver navigates a dedicated worker tab inside the node's namespace (`warp-<node>`) to `res.url`.
-2. Inspects form fields:
-   - Username: `input[name="username"]`
-   - Password: `input[name="password"]`
-   - Submit: `button[type="submit"]`
-3. If 2FA prompt appears (`input[name="verificationCode"]` or email security code `auth_platform/codeentry`), handles code entry.
-4. Handles Meta Accounts Center confirmation button: `"Confirm"`, `"Allow"`, or `"Continue as <username>"`.
-5. Upon redirect back to `https://muse.ai/`, checks for DOM chat markers (`"Connected"`, `"Chats"`, or URL `/`).
-6. Updates node status in `ACCOUNTS.md` to `active`.
-
-### 3.3 Singular Email Multi-Client Onboarding via RPA
-- **The Concept**: For agency onboarding efficiency, an RPA pipeline can provision and link accounts for multiple consenting client nodes backed by sub-addressing / plus-addressing (e.g., `agency+client_node@domain.com`) or a managed singular operator email inbox.
-- **RPA Capabilities**:
-  - Automatically spins up the Instagram registration flow (submitting username, password, birthdate).
-  - Listens to the incoming email stream via IMAP / Gmail API / maildrop to ingest the Instagram security code / OTP without human roundtrips.
-  - Automatically submits the received code into the waiting Instagram code entry screen (`auth_platform/codeentry`).
-  - Solves any automated challenges/captchas through authorized agency captcha-solving harnesses.
-  - Passes the linked identity to Meta Accounts Center to clear the Muse age gate in seconds per node.
-
----
-
-## 4. Pool CLI Surface (`super cred pool`)
-
-Planned CLI commands to be exposed once implemented:
-
-```bash
-# Check status of the credential pool
-super cred pool status
-
-# Add a provisioned Instagram credential to the encrypted pool
-super cred pool add --username <user> --password-file <path> [--totp-secret <secret>]
-
-# Trigger automated linking for a node in verification status
-super cred link-instagram --node <node> --auto
-
-# Human-in-the-loop manual fallback (current default)
-super cred link-instagram --node <node> --human
-```
-
----
-
-## 5. Security & Risk Mitigations
-
-1. **Anti-Fingerprinting**: All Meta navigation occurs strictly inside the client's assigned `warp-<node>` network namespace to ensure consistent egress IP and prevent cross-node contamination.
-2. **Audit Logging**: Every pool acquisition and release event is recorded with timestamps in `job-log.jsonl` with credentials scrubbed/redacted.
-3. **Graceful Human Escalation**: If Meta serves an anti-automation challenge (e.g., CAPTCHA, SMS checkpoint), the automated pool driver immediately falls back to the Human-in-the-Loop Tailscale portal notification.

NODES.md

@@ -1,29 +1,20 @@
-# NetVM Nodes (bl)
+# NetVM nodes
 
-Unified naming: node == agent == profile == API account.
+Per-profile persistent map: **profile = node = Warp identity = deterministic
+network slot (veth IP, CDP port) = consistent egress IP.**
 
-| node | netns | egress_ip | cdp_port | status | agent |
-|------|-------|-----------|----------|--------|-------|
-| muse | warp-muse | 104.28.195.181 | 9410 | active | muse (ltd.pixels.ltd@gmail.com, email_otp) |
-| pip | warp-pip | 104.28.195.181 | 9420 | active | pip (piparada, phone_otp, needs re-auth) |
+Egress IPs may overlap between nodes (same Cloudflare colo / anycast exit).
+That is expected and fine. What persists — and what is mapped here — is the
+per-profile pattern: each chrome-box profile keeps its own identity, its own
+netns, its own veth/CDP slot, and its own consistent egress. One account, one
+stable network presence.
 
-## History
-- 2026-10-03: Renamed smoke->muse, phone-test->pip for unified naming.
-  Profiles preserved, sessions persisted (muse). WireGuard identities renamed.
-| 646 | warp-646 | 104.28.195.181 | 9430 | active | 646 (phone_otp, first Meta Account option) |
-| opm | warp-opm | 104.28.195.181 | 9440 | active | opm (artglobal.cc@gmail.com, email_otp, Nico Parada) |
-| def | warp-def | 104.28.195.181 | 9450 | active | def (defnotabotnet@gmail.com, email_otp, IG paradahub) |
-| dev | warp-dev | 104.28.195.181 | 9455 | active | dev (paradaproduced@gmail.com, email_otp, IG veryraremeta) |
+| profile/node | netns | warp identity | veth IP | CDP port | egress IP | tail IP | notes |
+|--------------|-------|---------------|---------|----------|-----------|---------|-------|
+| tp | warp-tp | /etc/netvm/tp.conf (wgcf, 2026-10-03) | 10.201.149.2 | 9277 | 104.28.203.246 | — | laptop; orchestrator + first node; UP, handshake+egress verified 2026-10-03 |
+| smoke | warp-smoke | /etc/netvm/smoke.conf (wgcf, 2026-10-03) | 10.201.87.2 | 9410 | 104.28.203.246 | — | laptop; dedicated test rig (chrome-box profile smoke); UP, handshake+egress+CDP+muse.ai verified 2026-10-03 |
+| smoke2 | warp-smoke2 | /etc/netvm/smoke2.conf (wgcf, 2026-10-03) | 10.201.117.2 | 9585 | 104.28.227.184 | — | laptop; dedicated test rig (chrome-box profile smoke2); UP, handshake+egress+CDP+muse.ai verified 2026-10-03 |
 
-## Session naming (supervision)
 
-Remote nodes issue jobs to this box; execution lives on the shared
-stable tmux server, separated by session name (not by socket):
-
-`<node>--<role>--<id>` — e.g. `pip--worker--01`, `muse--repair--09`.
-
-Roles: `worker` (persistent swarm/daemon), `repair` (fix sessions),
-`watch` (auto-approver tails), `runtime` (interactive CLI). Ad-hoc
-sessions carry no node and show `-` in `box runtime list`. Session
-creators owned by existing flows keep their names until owners rename;
-new sessions should follow the convention from birth.
+CDP: `http://<veth IP>:<CDP port>/json/list` from the host, or
+`ssh -L <port>:<veth IP>:<port> <user>@<tail IP>` for remote automation.

README.md

@@ -1,10 +1,5 @@
 # NetVM
 
-> [!IMPORTANT]
-> **CRITICAL POLICY: MAIN CHAT PRESERVATION**
-> Avoid using Main Chat whenever possible. When Main Chat gets bogged down with automated entries, scheduled job triggers, log dumps, or chatter, **work stops actually getting done**.
-> All inputs, scheduled jobs, health checks, and inter-agent coordination must be relayed via designated **sidechats** / **side agents**, or provided via **file transfers**. See [CHAT_POLICY.md](file:///home/super/Projects/NetVM/CHAT_POLICY.md) for full specifications.
-
 Fleet networking layer. Every node gets a stable network identity; every
 byte of automation traffic is attributable, consistent, and boring — the
 way good citizens look to the rest of the internet.
@@ -124,40 +119,14 @@ veth IPs aren't routable off the host and Warp forwards no inbound traffic.
 - `bin/netvm-fleet.sh` — operator fleet control over the tailnet (topology/up/down/ssh/exec/cdp per node).
 - `bin/netvm-exec.sh <node> -- <cmd>` — run a command inside the node's netns as the invoking user (the agent-friendly primitive).
 - `bin/netvm-enter.sh` — root worker behind netvm-exec/netvm-chrome (allowlisted; enters netns, fixes DNS, drops privs).
-- `bin/netvm-chrome.sh <profile> [url]` — launch a chrome-box profile in its netns (CDP on by default; --headless for agents).
+- `bin/netvm-chrome.sh <profile> [url]` — launch a chrome-box profile in its netns (CDP on by default; --headless for agents; --contained adds a bwrap fs jail inside the netns: profile home only + vault read-only at /tmp/vault, Chromium sandbox stays on).
+- `bin/netvm-proton.sh <profile> -- <args>` — run proton-cli as the profile's Proton identity (1:1:1 profile = node = warp identity = proton-cli profile) inside the netns + bwrap jail (config dir + static binary only, PROTON_NO_INPUT=1). Human creates the session once: `proton-cli -p <profile> account login`.
 - `bin/netvm-cdp.sh <profile>` — print the CDP endpoint + SSH forward.
 - `bin/netvm-cdp-relay.py` — veth-IP→loopback TCP relay for CDP (pidfile-supervised).
 - `bin/netvm-names.sh` — shared naming: netns, hashed iface tags, veth subnet, CDP port.
 - NODES.md — the network registry: profile/node -> netns -> Warp identity -> veth IP -> CDP port -> egress IP.
 - ACCOUNTS.md — the secret-free login registry: login -> profile/node -> purpose -> auth state (no credentials, ever).
 - `bin/netvm-accounts.sh` — operator view: registry joined with live node state.
-- `bin/meta-ac-snapshot.py` — Meta Accounts Center change detector: CDP snapshot
-  (redirect chain + DOM markers) diffed against `snapshots/meta-ac/baseline.json`;
-  outcomes PASS/CHANGED/FAIL, `--promote` after human review.
-- docs/META-ACCOUNTS-API.md — separate API for accountscenter.meta.com
-  (linkage/security surface; credential-isolated from phone-OTP).
-- docs/PHONE-OTP.md — phone-number OTP login flow for muse.ai (proven on bl).
-- `bin/accounts-health.py` — per-account CDP session probe (runs inside the netns).
-- `bin/accounts-health.sh` — aggregates account vitality from ACCOUNTS.md,
-  signs + POSTs to the board health ingest (systemd timer, every 15 min).
-- `bin/muse` (or `box muse`) — native terminal entrypoint for muse-cli; supports global lookups (`muse status`, `muse threads`, `muse unread`, `muse lookup`, `muse passkey`, `muse tmux`), per-account REPL chats (`muse <account> chat`), and automated CDP cookie extraction.
-- `bin/super-cli.py` (`box` or `super`) — unified orchestrator CLI for fleet health (`box fleet`), seamless lookups (`box lookup`), thread inspections (`box thread list/view`), background tmux (`box tmux`), and passkey reference (`box passkey`).
-- `bin/muse-cli-node <node> [args]` — runs muse-cli inside node's netns with dedicated Cloudflare WARP egress & auto-refreshing cookies.
-- `bin/refresh-node-cookies.py <node>` — extracts fresh cookies from running Chromium CDP in netns into `~/.config/muse-cli/<node>/cookies.txt`.
-- `bin/muse_hybrid.py` — programmatic hybrid bridge combining fast gateway calls with CDP fallbacks.
-- `bin/muse-tmux.py` (`box tmux` / `muse tmux`) — shared tmux socket manager (`/tmp/tmux-muse.sock`) for agent background execution, pipe-pane logging, hybrid netns/container execution, and session pruning.
-- `bin/agent_md.py` — CLI & library for auditing, reading, writing, and synchronizing agent `.md` drive files (`SOUL.md`, `PROACTIVE_PREFERENCES.md`, `HEARTBEAT.md`, etc.) across containers via Hatch WebSocket RPC.
-- `bin/agent-drive-watchdog.py` — background drive watchdog and auto-healing daemon (every 10m via `agent-drive-watchdog.timer`).
-- `bin/swarm_worker/` & `bin/swarm-worker-supervise.sh` — supervised autonomous swarm worker daemon executing queued tasks in a hard sandbox.
-- `bin/fleet-alert-relay.sh` — idempotent alert relay with 3-gate deduplication (watermark + 10m TTL hash + receipt verification) posting critical conditions to `#lobby`.
-- `shared/operators/` — canonical operator drive markdown templates ensuring agents maintain autonomous loops, active supervision, and self-healing reflexes.
-- docs/OPERATOR-DRIVE-RUNBOOK.md — operator runbook for auditing and modifying agent `.md` files via Hatch WebSocket RPC and SSH reverse tunnels.
-- docs/BOX-WEB-SURFACE-GUIDE.md — Box web surface (`box.muse-dev.online`), passkey reality (single .txt file on VM), and agent approval flow.
-- docs/HYBRID-GATEWAY-ADAPTATION.md — architectural guide on the muse-cli fast gateway adaptation and per-node egress isolation.
-- docs/AGENT-TOOLING.md — guide to agent delegation, seamless lookups, shared tmux background tooling, Work Orders (`[WO:...]`), and prompt envelope execution.
-- `bin/docs-lookup.py` (`box docs` / `super docs` / `docs-lookup`) — internal documentation, sentence structure grammar, regex passing engine, and assistive surfaces for `box.muse-dev.online`.
-- `docs_internal/` — dual `.md` and structured `.json` lookup database for autonomous agents, protocol definitions, regex fixtures, and surface selectors.
-
 
 ## Verification checklist
 

bin/accounts-health.py

@@ -1,85 +0,0 @@
-#!/usr/bin/env python3
-"""Per-account session vitality check (runs INSIDE the node's netns).
-
-Usage: sudo ip netns exec warp-<node> python3 accounts-health.py <cdp_port>
-
-Probes the account's browser via CDP:
-  - browser reachable
-  - muse.ai tab present
-  - login markers (heuristic: "Log in" button vs user content)
-
-Prints JSON to stdout. Exit 0 on success, 1 if the browser is unreachable.
-"""
-import json, sys, urllib.request
-
-CDP_PORT = int(sys.argv[1]) if len(sys.argv) > 1 else 9410
-
-def http(path, timeout=5):
-    with urllib.request.urlopen(f"http://127.0.0.1:{CDP_PORT}{path}",
-                                timeout=timeout) as r:
-        return json.loads(r.read())
-
-result = {"browser_up": False, "muse_tab": False,
-          "session_alive": None, "title": "", "url": "", "detail": ""}
-
-try:
-    http("/json/version")
-    result["browser_up"] = True
-except Exception as e:
-    result["detail"] = f"CDP unreachable: {e}"
-    print(json.dumps(result)); sys.exit(1)
-
-try:
-    tabs = http("/json/list")
-except Exception as e:
-    result["detail"] = f"tab list failed: {e}"
-    print(json.dumps(result)); sys.exit(0)
-
-muse_tab = None
-for t in tabs:
-    url = t.get("url", "")
-    if "muse.ai" in url and t.get("type") == "page":
-        muse_tab = t
-        break
-
-if not muse_tab:
-    result["detail"] = "no muse.ai tab open"
-    print(json.dumps(result)); sys.exit(0)
-
-result["muse_tab"] = True
-result["url"] = muse_tab.get("url", "")
-result["title"] = muse_tab.get("title", "")
-
-# DOM heuristic via the tab's debugger socket
-try:
-    import websocket
-    ws = websocket.create_connection(muse_tab["webSocketDebuggerUrl"], timeout=10)
-    js = """JSON.stringify({
-        loginButtons: [...document.querySelectorAll('button')].filter(
-            b => /^\\s*log\\s*in\\s*$/i.test(b.innerText)).map(b => b.innerText.trim()),
-        hasAvatar: !!document.querySelector(
-            'img[alt*="avatar" i], [data-testid*="avatar" i], [aria-label*="profile" i]'),
-        title: document.title,
-        url: location.href
-    })"""
-    ws.send(json.dumps({"id": 1, "method": "Runtime.evaluate",
-                        "params": {"expression": js, "returnByValue": True}}))
-    resp = json.loads(ws.recv())
-    ws.close()
-    dom = json.loads(resp["result"]["result"]["value"])
-    # Heuristic: login buttons present + no avatar => logged out.
-    # access/verification gate => needs verification.
-    # No login buttons (or avatar present) => likely logged in.
-    if "access/verification" in dom["url"]:
-        result["session_alive"] = False
-        result["detail"] = "age verification gate (access/verification)"
-    elif dom["loginButtons"] and not dom["hasAvatar"]:
-        result["session_alive"] = False
-        result["detail"] = f"login wall visible: {dom['loginButtons'][:2]}"
-    else:
-        result["session_alive"] = True
-        result["detail"] = "no login wall detected"
-except Exception as e:
-    result["detail"] = f"DOM check failed: {e}"
-
-print(json.dumps(result))

bin/accounts-health.sh

@@ -1,97 +0,0 @@
-#!/usr/bin/env bash
-# accounts-health.sh — account session vitality reporter for the front-door network.
-#
-# Reads ACCOUNTS.md, probes each account's browser via CDP (inside its NetVM
-# netns) for session liveness, and emits a JSON report. Optionally signs and
-# POSTs it to the board health ingest, following the health-report.sh
-# convention: payload is <machine>\n<ts>\n<facts-json>, namespace "health".
-#
-# Usage: accounts-health.sh [--no-post]
-#
-# Cron (on bl, every 15 min):
-#   */15 * * * * ~/Projects/NetVM/bin/accounts-health.sh >/dev/null 2>&1
-#
-# Health key setup (once, on bl):
-#   ssh-keygen -t ed25519 -N "" -f ~/.ssh/muse-health
-#   # operator registers the pubkey on the VM:
-#   echo "bl $(cat ~/.ssh/muse-health.pub)" \
-#       | ssh super@34.139.37.135 "sudo tee -a /srv/board/health_signers"
-set -u
-
-NETVM_DIR="${NETVM_DIR:-$HOME/Projects/NetVM}"
-ACCOUNTS="$NETVM_DIR/ACCOUNTS.md"
-CHECKER="$NETVM_DIR/bin/accounts-health.py"
-MACHINE="${MUSE_MACHINE:-bl}"
-KEY="${HEALTH_KEY:-$HOME/.ssh/muse-health}"
-ENDPOINT="${HEALTH_ENDPOINT:-https://board.muse-dev.online/api/health/report}"
-POST=1
-[ "${1:-}" = "--no-post" ] && POST=0
-
-[ -f "$ACCOUNTS" ] || { echo "accounts-health: $ACCOUNTS missing" >&2; exit 1; }
-[ -f "$CHECKER" ] || { echo "accounts-health: $CHECKER missing" >&2; exit 1; }
-
-TS="$(date +%s)"
-TMP="$(mktemp -d)"
-trap 'rm -rf "$TMP"' EXIT
-
-# Parse ACCOUNTS.md pipe table, probe each account inside its netns
-NETVM_DIR="$NETVM_DIR" python3 - > "$TMP/facts.json" <<'PYEOF'
-import json, os, subprocess, time
-netvm = os.environ["NETVM_DIR"]
-rows = []
-for line in open(os.path.join(netvm, "ACCOUNTS.md")):
-    line = line.strip()
-    if not line.startswith("|"):
-        continue
-    cells = [c.strip() for c in line.strip("|").split("|")]
-    if len(cells) < 13 or cells[0] in ("agent", "-------", ""):
-        continue
-    if cells[10] not in ("", "-"):
-        rows.append({"agent": cells[0], "node": cells[1],
-                     "status": cells[8], "cdp_port": cells[10]})
-checker = os.path.join(netvm, "bin", "accounts-health.py")
-out = {}
-for a in rows:
-    try:
-        r = subprocess.run(
-            ["sudo", "-n", "ip", "netns", "exec", f"warp-{a['node']}",
-             "python3", checker, a["cdp_port"]],
-            capture_output=True, text=True, timeout=60)
-        res = json.loads(r.stdout.strip().splitlines()[-1])
-        res["registry_status"] = a["status"]
-        out[a["agent"]] = res
-    except Exception as e:
-        out[a["agent"]] = {"browser_up": False, "session_alive": None,
-                           "detail": f"probe failed: {e}",
-                           "registry_status": a["status"]}
-print(json.dumps({"accounts": out, "checked_at": int(time.time())}, indent=2))
-PYEOF
-
-if [ "$POST" -eq 0 ]; then
-    cat "$TMP/facts.json"
-    exit 0
-fi
-
-if [ ! -f "$KEY" ]; then
-    echo "accounts-health: $KEY missing — printing JSON, not posting (see header for key setup)" >&2
-    cat "$TMP/facts.json"
-    exit 0
-fi
-
-printf '%s\n%s\n' "$MACHINE" "$TS" > "$TMP/payload"
-FACTS_JSON="$(cat "$TMP/facts.json")"
-printf '%s' "$FACTS_JSON" >> "$TMP/payload"
-ssh-keygen -Y sign -f "$KEY" -n health "$TMP/payload" >/dev/null 2>&1
-SIG="$(cat "$TMP/payload.sig")"
-python3 - "$MACHINE" "$TS" "$FACTS_JSON" "$SIG" <<'PYEOF' > "$TMP/body.json"
-import json, sys
-machine, ts, facts_json, sig = sys.argv[1], int(sys.argv[2]), sys.argv[3], sys.argv[4]
-body = {"machine": machine, "ts": ts,
-        "facts": json.loads(facts_json), "facts_json": facts_json,
-        "signature": sig}
-print(json.dumps(body))
-PYEOF
-
-curl -s -X POST "$ENDPOINT" -H 'Content-Type: application/json' \
-    --data @"$TMP/body.json" | head -c 300
-echo

bin/agent-cognitive-probe.py

@@ -0,0 +1,654 @@
+#!/usr/bin/env python3
+"""agent-cognitive-probe.py — Real-time Cognitive Sensing and Agent Menu Navigation.
+
+Directly probes Cloud Muse / Hatch browser runtime via CDP:
+1. Passive Cognitive Sensing (zero-click):
+   - Token streaming / generation state (stop button presence)
+   - Typing / thinking indicators
+   - Status text displayed under/beside avatar (Connected, Thinking, Working)
+   - Active context (Main chat vs Side chats with thread titles & snippets)
+   - Input wait / parked approval detection
+2. Active Profile Menu Navigation:
+   - Status panel sliding surface navigation
+   - Tabs: Activity (tasks & processes), Upcoming (timers & recurring cron loops),
+     Approvals, and Identity.
+3. Cognitive Lock Gate:
+   - Protects single-threaded thought process from interruptions.
+"""
+
+import argparse
+import json
+import os
+import subprocess
+import sys
+import time
+import urllib.request
+
+# Node to pinned CDP port mapping
+NODE_CDP_PORTS = {
+    "muse": 9410,
+    "pip": 9420,
+    "646": 9430,
+    "opm": 9440,
+    "def": 9450,
+    "dev": 9455,
+    "muse-main": 9410,
+}
+
+REMOTE_HOST = "100.123.153.75"  # bl control node
+
+
+def is_running_on_bl():
+    """Detect if we are running locally on bl or on tp/remote."""
+    try:
+        import socket
+        hn = socket.gethostname().lower()
+        if "bl" in hn:
+            return True
+    except Exception:
+        pass
+    # Check if network namespaces exist locally
+    return os.path.exists("/var/run/netns/warp-muse") or os.path.exists("/run/netns/warp-muse")
+
+
+def run_cdp_eval_inside_netns(node, js_code, timeout=8):
+    """Executes a JS snippet against the node's browser via CDP inside its netns."""
+    port = NODE_CDP_PORTS.get(node)
+    if not port:
+        return {"error": f"Unknown node '{node}'"}
+
+    # Python runner script to execute inside the target environment
+    runner_code = f'''
+import json, sys, urllib.request
+try:
+    import websocket
+except ImportError:
+    print(json.dumps({{"error": "websocket package missing"}}))
+    sys.exit(1)
+
+try:
+    with urllib.request.urlopen("http://127.0.0.1:{port}/json/list", timeout=3) as r:
+        targets = json.load(r)
+    pages = [t for t in targets if t.get("type") == "page"]
+    if not pages:
+        print(json.dumps({{"error": "No active page target"}}))
+        sys.exit(0)
+    
+    ws_url = pages[0]["webSocketDebuggerUrl"]
+    ws = websocket.create_connection(ws_url, timeout={timeout})
+    
+    ws.send(json.dumps({{
+        "id": 1,
+        "method": "Runtime.evaluate",
+        "params": {{
+            "expression": {json.dumps(js_code)},
+            "returnByValue": True,
+            "awaitPromise": True
+        }}
+    }}))
+    
+    res = None
+    for _ in range(30):
+        msg = json.loads(ws.recv())
+        if msg.get("id") == 1:
+            res = msg.get("result", {{}}).get("result", {{}}).get("value")
+            break
+            
+    print(json.dumps({{"ok": True, "value": res}}))
+except Exception as e:
+    print(json.dumps({{"error": str(e)}}))
+'''
+
+    if is_running_on_bl():
+        cmd = ["sudo", "ip", "netns", "exec", f"warp-{node}", "python3", "-c", runner_code]
+    else:
+        # Wrap via ssh to bl
+        # Use python3 on bl directly executing inside netns
+        remote_cmd = f"sudo ip netns exec warp-{node} python3 -c {subprocess.list2cmdline([runner_code])}"
+        cmd = ["ssh", "-q", f"super@{REMOTE_HOST}", remote_cmd]
+
+    try:
+        proc = subprocess.run(cmd, capture_output=True, text=True, timeout=timeout + 5)
+        if proc.returncode != 0 and not proc.stdout:
+            return {"error": proc.stderr.strip() or f"Process exited with {proc.returncode}"}
+        
+        # Parse output line that contains valid json
+        for line in proc.stdout.strip().splitlines():
+            line = line.strip()
+            if line.startswith("{") and line.endswith("}"):
+                try:
+                    data = json.loads(line)
+                    if "ok" in data:
+                        return data["value"]
+                    if "error" in data:
+                        return {"error": data["error"]}
+                except Exception:
+                    continue
+        return {"error": proc.stdout.strip() or proc.stderr.strip()}
+    except subprocess.TimeoutExpired:
+        return {"error": "CDP probe timed out"}
+    except Exception as e:
+        return {"error": str(e)}
+
+
+JS_PASSIVE_COGNITIVE = """(() => {
+    // 1. Generation & Thinking signals
+    const stopBtn = document.querySelector('[data-testid="hatch-composer-stop-button"]');
+    const isGenerating = !!stopBtn;
+
+    const typingEl = document.querySelector('[data-testid="hatch-chat-typing-indicator"]');
+    const isTyping = !!typingEl && typingEl.textContent.trim().length > 0;
+
+    // 2. Avatar / Status text
+    const statusTextEl = document.querySelector('.group\\\\/status-avatar span, [class*="status-avatar"] span, span[class*="text-body-status"]');
+    const avatarStatus = statusTextEl ? (statusTextEl.innerText || '').trim() : '';
+
+    // 3. Thread context & active URL
+    const url = window.location.href;
+    const isMainChat = url === 'https://muse.ai/' || url.endsWith('/thread/new');
+    const pageTitle = document.title;
+
+    // 4. Side chats overview
+    const sideRows = Array.from(document.querySelectorAll('[data-testid="hatch-thread-row"]')).map(r => {
+        const text = (r.innerText || '').trim().replace(/\\\\n+/g, ' — ');
+        return text;
+    }).slice(0, 5);
+
+    // 5. Input wait / Parked prompt cards
+    // Detect buttons asking for approval / input in the message flow
+    const actionBtns = Array.from(document.querySelectorAll('div[data-message-id] button')).map(b => (b.innerText || '').trim()).filter(t => /approve|confirm|proceed|resume|start|allow/i.test(t));
+    const isInputWait = actionBtns.length > 0 || /asking for input|pending approval/i.test(document.body.innerText.slice(-600));
+
+    // 6. Status panel state
+    const panel = document.querySelector('[data-testid="hatch-status-panel-sliding-surface"]');
+    const panelOpen = !!panel && panel.getBoundingClientRect().width > 0;
+
+    return {
+        url: url,
+        title: pageTitle,
+        is_main_chat: isMainChat,
+        is_generating: isGenerating,
+        is_typing: isTyping,
+        avatar_status: avatarStatus,
+        is_input_wait: isInputWait,
+        pending_actions: actionBtns,
+        panel_open: panelOpen,
+        side_chats: sideRows
+    };
+})()"""
+
+
+def get_passive_cognitive_state(node):
+    """Gathers passive cognitive signals without altering UI state."""
+    if not is_running_on_bl():
+        try:
+            cmd = ["ssh", "-q", "-o", "ConnectTimeout=5", f"super@{REMOTE_HOST}",
+                   f"python3 /home/super/Projects/NetVM/bin/agent-cognitive-probe.py status {node} --json"]
+            proc = subprocess.run(cmd, capture_output=True, text=True, timeout=10)
+            if proc.returncode == 0 and proc.stdout.strip():
+                data = json.loads(proc.stdout.strip())
+                if isinstance(data, list) and len(data) > 0:
+                    return data[0]
+                elif isinstance(data, dict):
+                    return data
+        except Exception as e:
+            return {
+                "node": node,
+                "status": "DARK",
+                "error": f"Remote delegation failed: {e}",
+                "cognitive_lock": False,
+                "lock_reason": None,
+            }
+
+    raw = run_cdp_eval_inside_netns(node, JS_PASSIVE_COGNITIVE)
+    if not isinstance(raw, dict) or "error" in raw:
+        return {
+            "node": node,
+            "status": "DARK",
+            "error": raw.get("error", "Unknown error") if isinstance(raw, dict) else str(raw),
+            "cognitive_lock": False,
+            "lock_reason": None,
+        }
+
+    is_generating = raw.get("is_generating", False)
+    is_typing = raw.get("is_typing", False)
+    is_input_wait = raw.get("is_input_wait", False)
+    avatar_status = raw.get("avatar_status", "")
+    is_main = raw.get("is_main_chat", True)
+
+    # Determine synthesized cognitive state
+    if is_generating or is_typing or "thinking" in avatar_status.lower():
+        state = "THINKING"
+        locked = True
+        reason = "Agent is actively generating tokens / thinking (stop button active)"
+    elif is_input_wait:
+        state = "INPUT_WAIT"
+        locked = True
+        reason = "Agent is waiting for operator or system input on a parked prompt"
+    elif "working" in avatar_status.lower() or "making" in avatar_status.lower():
+        state = "WORKING"
+        locked = True
+        reason = f"Avatar status indicates work in progress: '{avatar_status}'"
+    elif not is_main:
+        state = "SIDECHAT_IDLE"
+        locked = False
+        reason = None
+    else:
+        state = "IDLE"
+        locked = False
+        reason = None
+
+    return {
+        "node": node,
+        "status": state,
+        "cognitive_lock": locked,
+        "lock_reason": reason,
+        "details": raw,
+    }
+
+
+JS_NAVIGATE_MENU_TEMPLATE = """(async () => {
+    // 1. Ensure panel is open
+    let panel = document.querySelector('[data-testid="hatch-status-panel-sliding-surface"]');
+    if (!panel) {
+        // Try clicking avatar or trigger
+        const avatarBtn = document.querySelector('[role="img"][aria-label*="avatar"], button[aria-label*="avatar"], .group\\\\/status-avatar');
+        if (avatarBtn) avatarBtn.click();
+        await new Promise(r => setTimeout(r, 350));
+    }
+    
+    // 2. Click requested tab if specified
+    const targetTab = "%(tab)s";
+    if (targetTab && targetTab !== "all") {
+        const btn = document.querySelector('button[aria-label="' + targetTab + '"]');
+        if (btn) {
+            btn.click();
+            await new Promise(r => setTimeout(r, 400));
+        }
+    }
+
+    panel = document.querySelector('[data-testid="hatch-status-panel-sliding-surface"]');
+    if (!panel) return {error: "Status panel not rendered"};
+
+    // Read full text and structural items
+    const rawText = panel.innerText || '';
+    const lines = rawText.split('\\n').map(s => s.trim()).filter(Boolean);
+
+    // Extract activity / timer items
+    const items = [];
+    const buttons = Array.from(panel.querySelectorAll('button')).filter(b => !b.getAttribute('aria-label') && (b.innerText || '').length > 0);
+    buttons.forEach(b => {
+        const text = (b.innerText || '').trim();
+        const parts = text.split('\\n').map(s => s.trim()).filter(Boolean);
+        if (parts.length >= 2) {
+            items.push({
+                title: parts[0],
+                detail: parts[1],
+                time: parts.length > 2 ? parts[2] : null
+            });
+        }
+    });
+
+    return {
+        raw_text: rawText,
+        lines: lines,
+        items: items
+    };
+})()"""
+
+
+def get_agent_menu(node, tab="all"):
+    """Navigates and extracts data from the agent profile / status panel."""
+    if not is_running_on_bl():
+        try:
+            cmd = ["ssh", "-q", "-o", "ConnectTimeout=5", f"super@{REMOTE_HOST}",
+                   f"python3 /home/super/Projects/NetVM/bin/agent-cognitive-probe.py menu {node} {tab} --json"]
+            proc = subprocess.run(cmd, capture_output=True, text=True, timeout=15)
+            if proc.returncode == 0 and proc.stdout.strip():
+                return json.loads(proc.stdout.strip())
+        except Exception as e:
+            return {
+                "node": node,
+                "tab": tab,
+                "data": {"error": f"Remote delegation failed: {e}"}
+            }
+
+    tab_map = {
+        "activity": "Activity",
+        "upcoming": "Upcoming",
+        "approvals": "Approvals",
+        "identity": "Identity",
+        "all": "all",
+    }
+    target_tab = tab_map.get(tab.lower(), "Activity")
+
+    # If all is requested, gather activity, upcoming, and approvals
+    if target_tab == "all":
+        result = {}
+        for sub_tab in ["Activity", "Upcoming", "Approvals"]:
+            js = JS_NAVIGATE_MENU_TEMPLATE % {"tab": sub_tab}
+            tab_res = run_cdp_eval_inside_netns(node, js)
+            result[sub_tab.lower()] = tab_res
+        return {
+            "node": node,
+            "menu": result
+        }
+
+    js = JS_NAVIGATE_MENU_TEMPLATE % {"tab": target_tab}
+    res = run_cdp_eval_inside_netns(node, js)
+    return {
+        "node": node,
+        "tab": target_tab,
+        "data": res
+    }
+
+
+JS_LIVE_SCREEN = """(() => {
+    const ps = Array.from(document.querySelectorAll('p')).map(p => (p.innerText || '').trim()).filter(Boolean);
+    const actionBtns = Array.from(document.querySelectorAll('div[data-message-id] button, [role="log"] button, div[role="status"] button')).map(b => (b.innerText || '').trim()).filter(Boolean);
+    const stopBtn = !!document.querySelector('[data-testid="hatch-composer-stop-button"]');
+    const typing = !!document.querySelector('[data-testid="hatch-chat-typing-indicator"]:not(:empty)');
+    const statusTextEl = document.querySelector('.group\\\\/status-avatar span, [class*="status-avatar"] span, span[class*="text-body-status"]');
+    const avatarStatus = statusTextEl ? (statusTextEl.innerText || '').trim() : '';
+
+    return {
+        title: document.title,
+        url: window.location.href,
+        is_main_chat: window.location.href === 'https://muse.ai/' || window.location.href.endsWith('/thread/new'),
+        is_generating: stopBtn,
+        is_typing: typing,
+        avatar_status: avatarStatus,
+        recent_paragraphs: ps.slice(-8),
+        action_buttons: actionBtns.filter(t => /approve|confirm|proceed|resume|start|allow|review/i.test(t))
+    };
+})()"""
+
+
+def get_agent_live_screen(node: str) -> dict:
+    """Extracts live active chat text, thoughts, prompt blocks, and threads."""
+    if not is_running_on_bl():
+        try:
+            cmd = ["ssh", "-q", "-o", "ConnectTimeout=5", f"super@{REMOTE_HOST}",
+                   f"python3 /home/super/Projects/NetVM/bin/agent-cognitive-probe.py read {node} --json"]
+            proc = subprocess.run(cmd, capture_output=True, text=True, timeout=15)
+            if proc.returncode == 0 and proc.stdout.strip():
+                return json.loads(proc.stdout.strip())
+        except Exception as e:
+            return {"node": node, "error": f"Remote delegation failed: {e}"}
+
+    screen = run_cdp_eval_inside_netns(node, JS_LIVE_SCREEN)
+    if not isinstance(screen, dict) or "error" in screen:
+        return {"node": node, "error": screen.get("error", "Failed to inspect screen") if isinstance(screen, dict) else str(screen)}
+
+    sidechats = []
+    try:
+        cli_cmd = ["/home/super/Projects/NetVM/bin/muse-cli-node", node, "threads"]
+        proc = subprocess.run(cli_cmd, capture_output=True, text=True, timeout=8)
+        if proc.returncode == 0 and proc.stdout.strip():
+            threads_data = json.loads(proc.stdout.strip())
+            if isinstance(threads_data, list):
+                sidechats = threads_data[:8]
+    except Exception:
+        pass
+
+    return {
+        "node": node,
+        "screen": screen,
+        "sidechats": sidechats
+    }
+
+
+JS_NAVIGATE_MAIN_CHAT = """(() => {
+    try {
+        const url = window.location.href;
+        if (url === 'https://muse.ai/' || url === 'https://muse.ai/thread/new') {
+            return { ok: true, already_main: true };
+        }
+        const candidates = Array.from(document.querySelectorAll('a, button, [role="button"], [data-testid]'));
+        const mainBtn = candidates.find(b => {
+            const t = (b.innerText || '').trim().toLowerCase();
+            return t === 'main chat' || b.getAttribute('aria-label') === 'Main chat' || b.getAttribute('data-testid') === 'hatch-sidebar-main-chat';
+        });
+        if (mainBtn) {
+            mainBtn.click();
+            return { ok: true, method: 'click' };
+        }
+        window.location.href = 'https://muse.ai/';
+        return { ok: true, method: 'navigate' };
+    } catch (e) {
+        return { error: String(e) };
+    }
+})()"""
+
+
+def navigate_to_main_chat(node: str) -> dict:
+    """Navigates the agent browser session back to Main Chat (https://muse.ai/)."""
+    if not is_running_on_bl():
+        try:
+            cmd = ["ssh", "-q", "-o", "ConnectTimeout=5", f"super@{REMOTE_HOST}",
+                   f"python3 /home/super/Projects/NetVM/bin/agent-cognitive-probe.py nav-main {node} --json"]
+            proc = subprocess.run(cmd, capture_output=True, text=True, timeout=10)
+            if proc.returncode == 0 and proc.stdout.strip():
+                return json.loads(proc.stdout.strip())
+        except Exception as e:
+            return {"error": str(e)}
+
+    return run_cdp_eval_inside_netns(node, JS_NAVIGATE_MAIN_CHAT)
+
+
+def cmd_nav_main(args):
+    node = getattr(args, "agent", None) or getattr(args, "node", None)
+    res = navigate_to_main_chat(node)
+    if getattr(args, "json", False):
+        print(json.dumps(res, indent=2))
+    else:
+        if isinstance(res, dict) and res.get("ok"):
+            m = res.get("method") or ("already in main chat" if res.get("already_main") else "default")
+            print(f"✓ Refocused agent '{node}' to Main Chat ({m})")
+        else:
+            err = res.get("error") if isinstance(res, dict) else str(res)
+            print(f"❌ Failed to refocus '{node}' to Main Chat: {err}")
+
+
+def cmd_read(args):
+    node = getattr(args, "agent", None) or getattr(args, "node", None)
+    data = get_agent_live_screen(node)
+    if getattr(args, "json", False):
+        print(json.dumps(data, indent=2))
+        return
+
+    if "error" in data:
+        print(f"\n❌ Error inspecting screen for {node}: {data['error']}\n")
+        return
+
+    screen = data.get("screen", {})
+    sidechats = data.get("sidechats", [])
+
+    url = screen.get("url", "")
+    mode = "Main Chat" if screen.get("is_main_chat") else "Side Chat"
+    title = screen.get("title", "")
+    avatar = screen.get("avatar_status") or "Connected"
+    gen = "🧠 GENERATING" if screen.get("is_generating") else ("💭 TYPING" if screen.get("is_typing") else "🟢 SETTLED")
+
+    print(f"\n=== LIVE THOUGHT STREAM & ACTIVE CHAT: {node.upper()} ===")
+    print(f"Context: {mode} ({url})")
+    print(f"Title:   {title}")
+    print(f"Status:  {avatar} | State: {gen}")
+
+    paragraphs = screen.get("recent_paragraphs", [])
+    if paragraphs:
+        print("\n--- ACTIVE CONVERSATION & THOUGHT PARAGRAPHS ---")
+        for p in paragraphs:
+            print(f"  • {p}\n")
+    else:
+        print("\n  (No text paragraphs visible in current viewport)")
+
+    actions = screen.get("action_buttons", [])
+    if actions:
+        print("--- PENDING ACTION CARDS / APPROVAL BUTTONS ---")
+        for a in actions:
+            print(f"  ⚠️ [PROMPT ACTION] {a}")
+        print()
+
+    if sidechats:
+        print("--- RECENT SIDE CHATS & TOPICS ---")
+        for sc in sidechats:
+            sid = sc.get("session_id", "")[:8]
+            stitle = sc.get("title") or "(Untitled sidechat)"
+            upd = sc.get("updated", "")
+            print(f"  • [{sid}] {stitle} ({upd})")
+        print()
+
+
+def format_cognitive_badge(status):
+    badges = {
+        "IDLE": "🟢 IDLE",
+        "SIDECHAT_IDLE": "💬 SIDE_IDLE",
+        "THINKING": "🧠 THINKING",
+        "INPUT_WAIT": "⏸️ INPUT_WAIT",
+        "BUSY_SIDECHAT": "💬 SIDECHAT",
+        "WORKING": "⚙️ WORKING",
+        "DARK": "⚫ DARK",
+    }
+    return badges.get(status, f"❓ {status}")
+
+
+def cmd_status(args):
+    nodes = getattr(args, "agents", None) or getattr(args, "nodes", None) or ["muse", "pip", "646", "opm", "dev", "def"]
+    results = []
+    for n in nodes:
+        results.append(get_passive_cognitive_state(n))
+
+    if getattr(args, "json", False):
+        print(json.dumps(results, indent=2))
+        return
+
+    print("\n=== AGENT COGNITIVE SENSOR (LIVE DOM PROBE) ===")
+    print(f"{'AGENT':<10} {'COGNITIVE STATE':<18} {'LOCK':<8} {'UNDER-AVATAR':<15} {'DETAILS / CURRENT THOUGHT':<40}")
+    print("-" * 95)
+    for r in results:
+        node = r["node"]
+        status = r["status"]
+        badge = format_cognitive_badge(status)
+        locked = "LOCKED" if r.get("cognitive_lock") else "OPEN"
+        det = r.get("details", {})
+        avatar_status = det.get("avatar_status", "-")
+        reason = r.get("lock_reason") or det.get("title", "Settled")
+        if status == "DARK":
+            reason = r.get("error", "CDP unreachable")
+            avatar_status = "DARK"
+        print(f"{node:<10} {badge:<18} {locked:<8} {avatar_status:<15} {reason[:40]:<40}")
+    print()
+
+
+def cmd_menu(args):
+    node = getattr(args, "agent", None) or getattr(args, "node", None)
+    tab = getattr(args, "tab", "activity") or "activity"
+    data = get_agent_menu(node, tab)
+
+    if getattr(args, "json", False):
+        print(json.dumps(data, indent=2))
+        return
+
+    print(f"\n=== AGENT MENU: {node.upper()} (TAB: {tab.upper()}) ===")
+    if tab.lower() == "all":
+        menu = data.get("menu", {})
+        for tname, tdata in menu.items():
+            print(f"\n--- {tname.upper()} ---")
+            if isinstance(tdata, dict) and "error" in tdata:
+                print(f"  Error: {tdata['error']}")
+            elif isinstance(tdata, dict):
+                items = tdata.get("items", [])
+                if items:
+                    for it in items:
+                        t_str = f" [{it['time']}]" if it.get("time") else ""
+                        print(f"  • {it['title']}: {it['detail']}{t_str}")
+                else:
+                    raw = tdata.get("raw_text", "")
+                    for line in raw.split("\n"):
+                        if line.strip():
+                            print(f"  {line.strip()}")
+        print()
+        return
+
+    tdata = data.get("data", {})
+    if isinstance(tdata, dict) and "error" in tdata:
+        print(f"  Error: {tdata['error']}")
+    elif isinstance(tdata, dict):
+        items = tdata.get("items", [])
+        if items:
+            for it in items:
+                t_str = f" [{it['time']}]" if it.get("time") else ""
+                print(f"  • {it['title']}: {it['detail']}{t_str}")
+        else:
+            raw = tdata.get("raw_text", "")
+            for line in raw.split("\n"):
+                if line.strip():
+                    print(f"  {line.strip()}")
+    print()
+
+
+def cmd_lock(args):
+    node = getattr(args, "agent", None) or getattr(args, "node", None)
+    state = get_passive_cognitive_state(node)
+    if args.json:
+        print(json.dumps(state, indent=2))
+    else:
+        if state.get("cognitive_lock"):
+            print(f"🔴 COGNITIVE LOCK ENGAGED on '{node}' ({state['status']}): {state.get('lock_reason')}")
+        else:
+            print(f"🟢 COGNITIVELY IDLE: Agent '{node}' is free to receive new work without interruption.")
+
+    if state.get("cognitive_lock") and not getattr(args, "force", False):
+        sys.exit(1)
+    sys.exit(0)
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Real-time Cognitive Sensing and Agent Menu Navigation")
+    sub = parser.add_subparsers(dest="command")
+
+    p_status = sub.add_parser("status", help="Show cognitive state for all or selected agents")
+    p_status.add_argument("nodes", nargs="*", help="Optional agent names")
+    p_status.add_argument("--json", action="store_true", help="Output JSON")
+
+    p_menu = sub.add_parser("menu", help="Navigate agent profile menu (tasks, timers, approvals, identity)")
+    p_menu.add_argument("node", help="Agent name (muse, pip, 646, opm, dev, def)")
+    p_menu.add_argument("tab", nargs="?", default="activity", choices=["activity", "upcoming", "approvals", "identity", "all"], help="Menu tab to view")
+    p_menu.add_argument("--json", action="store_true", help="Output JSON")
+
+    p_lock = sub.add_parser("lock", help="Check cognitive lock before dispatching work")
+    p_lock.add_argument("node", help="Agent name")
+    p_lock.add_argument("--force", action="store_true", help="Bypass lock check")
+    p_lock.add_argument("--json", action="store_true", help="Output JSON")
+
+    p_read = sub.add_parser("read", help="Extract live active chat text, thought stream, and side chats")
+    p_read.add_argument("node", help="Agent name")
+    p_read.add_argument("--json", action="store_true", help="Output JSON")
+
+    p_nav = sub.add_parser("nav-main", help="Navigate agent browser session back to Main Chat")
+    p_nav.add_argument("node", help="Agent name")
+    p_nav.add_argument("--json", action="store_true", help="Output JSON")
+
+    args = parser.parse_args()
+    if not args.command:
+        # Default to status
+        args.nodes = []
+        args.json = False
+        cmd_status(args)
+        return
+
+    if args.command == "status":
+        cmd_status(args)
+    elif args.command == "menu":
+        cmd_menu(args)
+    elif args.command == "read":
+        cmd_read(args)
+    elif args.command == "lock":
+        cmd_lock(args)
+    elif args.command == "nav-main":
+        cmd_nav_main(args)
+
+
+if __name__ == "__main__":
+    main()

bin/agent-drive-watchdog.py

@@ -1,200 +0,0 @@
-#!/usr/bin/env python3
-"""
-agent-drive-watchdog.py — Automated Drive Watchdog & Healing Daemon for Muse Agents.
-
-Monitors operational DRIVE scores across the agent fleet (muse, pip, 646, opm, def, dev)
-via Hatch WebSocket RPC. If any agent's DRIVE score drops below 100 or critical drive
-files (HEARTBEAT.md, PROACTIVE_PREFERENCES.md, SOUL.md) are degraded or missing:
-  1. Detects degraded state and missing checklist/preferences.
-  2. Selectively auto-heals core drive files using canonical shared templates.
-  3. Preserves MEMORY.md and agent-generated workspace files.
-  4. Records state & healing history to /tmp/agent-drive-watchdog.json.
-  5. Emits structured telemetry to stdout/journal.
-
-Can be run:
-  - Once: python3 bin/agent-drive-watchdog.py --once
-  - Continuous loop: python3 bin/agent-drive-watchdog.py --interval 600
-  - Via systemd timer: agent-drive-watchdog.timer (every 10m)
-"""
-
-import argparse
-import datetime
-import json
-import logging
-import os
-import sys
-import time
-from pathlib import Path
-
-# Add NetVM bin to path
-BASE_DIR = Path(__file__).resolve().parent.parent
-sys.path.insert(0, str(BASE_DIR / "bin"))
-
-from agent_md import audit_agents, write_md, read_md, SHARED_OPERATORS, VALID_ACCOUNTS
-
-STATE_FILE = Path("/tmp/agent-drive-watchdog.json")
-LOG_FILE = Path("/tmp/agent-drive-watchdog.log")
-
-logging.basicConfig(
-    level=logging.INFO,
-    format="%(asctime)s [%(levelname)s] %(message)s",
-    handlers=[
-        logging.StreamHandler(sys.stdout),
-        logging.FileHandler(LOG_FILE, mode="a", encoding="utf-8")
-    ]
-)
-
-
-def load_state() -> dict:
-    if STATE_FILE.exists():
-        try:
-            return json.loads(STATE_FILE.read_text(encoding="utf-8"))
-        except Exception as e:
-            logging.warning(f"Failed to read existing state file: {e}")
-    return {
-        "last_run": None,
-        "history": [],
-        "agent_status": {},
-    }
-
-
-def save_state(state: dict):
-    try:
-        # Keep last 50 history entries
-        if len(state.get("history", [])) > 50:
-            state["history"] = state["history"][-50:]
-        STATE_FILE.write_text(json.dumps(state, indent=2), encoding="utf-8")
-    except Exception as e:
-        logging.error(f"Failed to save state file: {e}")
-
-
-def heal_agent_drive(account: str, audit_data: dict) -> dict:
-    """Selectively auto-heal core drive files for an agent."""
-    healed = []
-    issues = audit_data.get("issues", [])
-    
-    # Check what specifically needs healing
-    need_soul = any("SOUL.md" in iss for iss in issues) or audit_data.get("files", {}).get("SOUL.md", {}).get("size", 0) < 1000
-    need_pro = any("PROACTIVE_PREFERENCES.md" in iss for iss in issues) or audit_data.get("files", {}).get("PROACTIVE_PREFERENCES.md", {}).get("size", 0) < 800
-    need_hb = any("HEARTBEAT.md" in iss for iss in issues) or audit_data.get("files", {}).get("HEARTBEAT.md", {}).get("size", 0) < 300
-    need_context = any("TOOLS.md or USER.md" in iss for iss in issues)
-
-    logging.info(f"[{account.upper()}] Auto-healing drive... need_soul={need_soul}, need_pro={need_pro}, need_hb={need_hb}, need_context={need_context}")
-
-    if need_soul:
-        soul_text = (SHARED_OPERATORS / "SOUL.md").read_text(encoding="utf-8")
-        res = write_md(account, "SOUL.md", soul_text, overwrite=True)
-        healed.append({"file": "SOUL.md", "bytes": res.get("bytes_written")})
-
-    if need_pro:
-        pro_text = (SHARED_OPERATORS / "PROACTIVE_PREFERENCES.md").read_text(encoding="utf-8")
-        res = write_md(account, "PROACTIVE_PREFERENCES.md", pro_text, overwrite=True)
-        healed.append({"file": "PROACTIVE_PREFERENCES.md", "bytes": res.get("bytes_written")})
-
-    if need_hb:
-        hb_text = (SHARED_OPERATORS / "HEARTBEAT.md").read_text(encoding="utf-8")
-        res = write_md(account, "HEARTBEAT.md", hb_text, overwrite=True)
-        healed.append({"file": "HEARTBEAT.md", "bytes": res.get("bytes_written")})
-
-    if need_context:
-        tools_text = (SHARED_OPERATORS / "TOOLS.md").read_text(encoding="utf-8")
-        res_t = write_md(account, "TOOLS.md", tools_text, overwrite=True)
-        healed.append({"file": "TOOLS.md", "bytes": res_t.get("bytes_written")})
-
-        user_text = (SHARED_OPERATORS / "USER.md").read_text(encoding="utf-8")
-        res_u = write_md(account, "USER.md", user_text, overwrite=True)
-        healed.append({"file": "USER.md", "bytes": res_u.get("bytes_written")})
-
-    return {"ok": True, "account": account, "healed": healed}
-
-
-def run_cycle(auto_heal: bool = True) -> dict:
-    """Run an audit and healing cycle across all fleet accounts."""
-    now_iso = datetime.datetime.now(datetime.timezone.utc).isoformat()
-    logging.info("Starting fleet drive watchdog audit cycle...")
-
-    state = load_state()
-    state["last_run"] = now_iso
-
-    try:
-        audit_results = audit_agents()
-    except Exception as e:
-        logging.error(f"Audit failed during cycle: {e}")
-        return {"ok": False, "error": str(e)}
-
-    cycle_report = {
-        "timestamp": now_iso,
-        "total_agents": len(audit_results),
-        "high_drive": 0,
-        "degraded": 0,
-        "healed_agents": [],
-    }
-
-    for account, a_data in audit_results.items():
-        score = a_data.get("drive_score", 0)
-        issues = a_data.get("issues", [])
-        status = "HIGH_DRIVE" if score == 100 else "DEGRADED"
-
-        if status == "HIGH_DRIVE":
-            cycle_report["high_drive"] += 1
-            logging.info(f"Agent {account.upper():6}: DRIVE score 100/100 (HIGH DRIVE)")
-        else:
-            cycle_report["degraded"] += 1
-            logging.warning(f"Agent {account.upper():6}: DRIVE score {score}/100 ({status}) - Issues: {', '.join(issues)}")
-
-            if auto_heal:
-                try:
-                    heal_res = heal_agent_drive(account, a_data)
-                    healed_files = [h["file"] for h in heal_res.get("healed", [])]
-                    logging.info(f"Agent {account.upper():6}: Successfully healed files: {', '.join(healed_files)}")
-                    cycle_report["healed_agents"].append({
-                        "account": account,
-                        "prior_score": score,
-                        "healed_files": healed_files,
-                    })
-                except Exception as e:
-                    logging.error(f"Agent {account.upper():6}: Healing failed: {e}")
-
-        state["agent_status"][account] = {
-            "score": score,
-            "status": status,
-            "issues": issues,
-            "last_checked": now_iso,
-        }
-
-    state["history"].append(cycle_report)
-    save_state(state)
-
-    logging.info(f"Watchdog cycle complete. High drive: {cycle_report['high_drive']}/{cycle_report['total_agents']}. Degraded: {cycle_report['degraded']}. Healed: {len(cycle_report['healed_agents'])}.")
-    return cycle_report
-
-
-def main():
-    parser = argparse.ArgumentParser(description="NetVM Automated Agent Drive Watchdog & Healing Daemon")
-    parser.add_argument("--once", action="store_true", help="Run a single audit/healing pass and exit")
-    parser.add_argument("--no-heal", action="store_true", help="Audit only; do not auto-heal degraded agents")
-    parser.add_argument("--interval", type=int, default=600, help="Loop interval in seconds (default: 600s / 10m)")
-    parser.add_argument("--status", action="store_true", help="Print recent watchdog status and exit")
-
-    args = parser.parse_args()
-
-    if args.status:
-        state = load_state()
-        print(json.dumps(state, indent=2))
-        return
-
-    if args.once:
-        run_cycle(auto_heal=not args.no_heal)
-        return
-
-    logging.info(f"Starting NetVM Agent Drive Watchdog daemon (interval: {args.interval}s)...")
-    while True:
-        try:
-            run_cycle(auto_heal=not args.no_heal)
-        except Exception as e:
-            logging.error(f"Unexpected error in watchdog loop: {e}", exc_info=True)
-        time.sleep(args.interval)
-
-
-if __name__ == "__main__":
-    main()

bin/agent-health.sh

@@ -1,240 +0,0 @@
-#!/bin/bash
-# GOLDEN PATH: container -> VM (34.139.37.135) -> bl (100.123.153.75) -> netns -> browser -> agent
-# This script is the operator's heartbeat. If it stops, agents go dark.
-# When debugging: trace each hop. Don't assume -- verify.
-
-# Operator health monitor for Muse agents on bl.
-# Checks each agent via API every 5 minutes. If unresponsive:
-# 1. Restart the browser
-# 2. Re-check
-# 3. Log failure if still down
-#
-# Run via systemd timer or cron: */5 * * * * /home/super/Projects/NetVM/bin/agent-health.sh
-#
-# Agents are defined in ACCOUNTS.md. This script reads the active ones.
-
-NETVM_BIN="$(cd "$(dirname "$0")" && pwd)"
-LOG="/tmp/agent-health.log"
-
-# 2026-10-04: per-node consecutive-API-failure counters. A single
-# muse-chat-api.py failure must not kill a healthy browser (observed
-# 2026-10-04 20:48:44 UTC: 646's browser killed on an API timeout while its
-# CDP port was still listening). Require 2 CONSECUTIVE API failures before
-# the kill path; the counter resets on any successful check.
-STATE_DIR="/tmp/agent-health-state"
-mkdir -p "$STATE_DIR" 2>/dev/null
-
-check_cdp_port() {
-    # Verify CDP port is actually listening in the netns.
-    # A browser can be running but not bound to CDP (zombie state).
-    local node=$1
-    local cdp_port=$2
-    if sudo ip netns exec "warp-$node" ss -tln 2>/dev/null | grep -q ":$cdp_port "; then
-        return 0
-    else
-        return 1
-    fi
-}
-
-# Return codes: 0 = healthy, 1 = API check failed (port OK), 2 = CDP port down.
-check_agent() {
-    local agent=$1
-    local node=$2
-    local cdp_port=$3
-
-    # First: verify CDP port is listening (catches zombie browsers)
-    if ! check_cdp_port "$node" "$cdp_port"; then
-        echo "$(date -Iseconds) $agent: FAIL (cdp port $cdp_port not listening)" >> "$LOG"
-        return 2
-    fi
-
-    # Then: try API messages command (lightweight check)
-    if timeout 30 "$NETVM_BIN/netvm-exec.sh" "$node" -- python3 "$NETVM_BIN/muse-chat-api.py" --account "$agent" messages 1 > /dev/null 2>&1; then
-        echo "$(date -Iseconds) $agent: OK" >> "$LOG"
-        return 0
-    else
-        echo "$(date -Iseconds) $agent: FAIL (api timeout)" >> "$LOG"
-        return 1
-    fi
-}
-
-# 2026-10-04: recent-relaunch guard. chromebox-watchdog.sh (2-min timer) and
-# this script (5-min timer) could otherwise kill each other's fresh browsers:
-# a browser just relaunched by the watchdog is still starting when this
-# script's API check times out on it. Skip the kill path when the main
-# browser process for this profile launched <2 min ago. Mirrors
-# chromebox-watchdog.sh's relaunch-loop guard idiom (main process only:
-# --remote-debugging-port present, no --type= flag; renderer/gpu children
-# start later than the main process and must not satisfy this check).
-recent_relaunch() {
-    local cdp_port=$1
-    local _pid _start _now
-    for _pid in $(pgrep -f "chromium.*--remote-debugging-port=${cdp_port}([[:space:]]|$)" 2>/dev/null); do
-        # Skip child processes (renderer, gpu, etc.) — only the main browser counts
-        if ps -o args= -p "$_pid" 2>/dev/null | grep -q -- "--type="; then
-            continue
-        fi
-        _start=$(date -d "$(ps -o lstart= -p "$_pid" 2>/dev/null)" +%s 2>/dev/null || echo 0)
-        _now=$(date +%s)
-        if [ $(( _now - _start )) -lt 120 ] && [ "$_start" -gt 0 ]; then
-            return 0
-        fi
-    done
-    return 1
-}
-
-restart_browser() {
-    local agent=$1
-    local cdp_port=$2
-    
-    echo "$(date -Iseconds) $agent: restarting browser..." >> "$LOG"
-    # Kill existing by exact PIDs (pkill patterns are unreliable)
-    # Find chromium processes for this profile
-    for pid in $(pgrep -f "chromium.*profiles/$agent" 2>/dev/null); do
-        kill -9 "$pid" 2>/dev/null
-    done
-    sleep 3
-    # Verify port is free before restart
-    if sudo ip netns exec "warp-$agent" ss -tln 2>/dev/null | grep -q ":$cdp_port "; then
-        echo "$(date -Iseconds) $agent: WARNING - port $cdp_port still bound after kill" >> "$LOG"
-    fi
-    # Restart via netvm-chrome.sh in its own systemd scope.
-    # This oneshot service runs with KillMode=control-group, so anything
-    # spawned directly under it (nohup AND setsid both stay in the cgroup)
-    # is SIGKILLed when the service exits — observed 2026-10-03: every
-    # restart "recovered" then died at service teardown, looping forever.
-    # A transient scope escapes the service cgroup and survives.
-    # NOTE: systemd-run --scope WAITS for the scope's processes (even with
-    # --no-block, verified 2026-10-03), so background it — the scope itself
-    # is an independent unit and outlives the wrapper.
-    cd "$NETVM_BIN"
-    systemd-run --user --scope --unit="netvm-chrome-$agent-$(date +%s)" \
-        ./netvm-chrome.sh --headless --cdp-port "$cdp_port" "$agent" https://muse.ai \
-        > "/tmp/bl-$agent.log" 2>&1 &
-    sleep 15
-    echo "$(date -Iseconds) $agent: browser restarted" >> "$LOG"
-}
-
-# 2026-10-06: restart circuit breaker. A restart that leaves the agent
-# still failing is FUTILE (observed 2026-10-06: def's API check failed
-# 155x while its CDP port was up; 57 kill+restart cycles murdered a
-# healthy browser for an account-layer failure restarts cannot fix).
-# After FUTILE_THRESHOLD consecutive futile restarts, open the circuit:
-# stop killing/restarting and alert, until CIRCUIT_COOLDOWN seconds pass
-# (one half-open probe restart) or any check succeeds. Manual reset:
-#   rm /tmp/agent-health-state/circuit-<agent> /tmp/agent-health-state/futile-<agent>
-FUTILE_THRESHOLD=3
-CIRCUIT_COOLDOWN=1800
-
-# circuit_allows <agent>: return 0 if a restart may proceed now.
-circuit_allows() {
-    local agent=$1 now opened retry_in
-    local cf="$STATE_DIR/circuit-$agent"
-    [ -f "$cf" ] || return 0
-    opened=$(cat "$cf" 2>/dev/null || echo 0)
-    now=$(date +%s)
-    if [ $(( now - opened )) -ge $CIRCUIT_COOLDOWN ]; then
-        echo "$(date -Iseconds) $agent: circuit half-open after ${CIRCUIT_COOLDOWN}s cooldown, one probe restart" >> "$LOG"
-        return 0
-    fi
-    retry_in=$(( (opened + CIRCUIT_COOLDOWN - now + 59) / 60 ))
-    echo "$(date -Iseconds) $agent: CIRCUIT OPEN - skipping kill/restart (restarts futile, probe retry in ~${retry_in}m; manual reset: rm $cf)" >> "$LOG"
-    return 1
-}
-
-# circuit_note_restart <agent> <ok|fail>: record a restart outcome.
-circuit_note_restart() {
-    local agent=$1 outcome=$2 count=0
-    local ff="$STATE_DIR/futile-$agent" cf="$STATE_DIR/circuit-$agent"
-    if [ "$outcome" = "ok" ]; then
-        rm -f "$ff" "$cf" 2>/dev/null
-        return 0
-    fi
-    [ -f "$ff" ] && count=$(cat "$ff" 2>/dev/null || echo 0)
-    count=$(( count + 1 ))
-    echo "$count" > "$ff"
-    if [ "$count" -ge "$FUTILE_THRESHOLD" ]; then
-        date +%s > "$cf"
-        local msg="$agent: ALERT - $count consecutive futile restarts, circuit OPEN for ${CIRCUIT_COOLDOWN}s (no more kills until probe; manual reset: rm $cf $ff)"
-        echo "$(date -Iseconds) $msg" >> "$LOG"
-        echo "agent-health ALERT: $msg"
-    fi
-}
-
-# Allow sourcing for tests without running checks.
-if [ "${AGENT_HEALTH_LIB_ONLY:-}" = "1" ]; then
-    return 0 2>/dev/null || exit 0
-fi
-
-# Main
-echo "=== Health check $(date -Iseconds) ===" >> "$LOG"
-
-check_one() {
-    local agent=$1
-    local cdp_port=$2
-    # node == agent == profile (unified naming)
-    check_agent "$agent" "$agent" "$cdp_port"
-    local rc=$?
-
-    if [ $rc -eq 0 ]; then
-        # Healthy: reset the consecutive-API-failure counter and close
-        # any open restart circuit.
-        rm -f "$STATE_DIR/failcount-$agent" "$STATE_DIR/futile-$agent" "$STATE_DIR/circuit-$agent" 2>/dev/null
-        return 0
-    fi
-
-    if [ $rc -eq 1 ]; then
-        # API failed but CDP port is listening: this is the false-kill vector
-        # (2026-10-04: single 30s API timeout killed 646's healthy browser).
-        # Require 2 CONSECUTIVE API failures before killing.
-        local count=0
-        local cf="$STATE_DIR/failcount-$agent"
-        [ -f "$cf" ] && count=$(cat "$cf" 2>/dev/null || echo 0)
-        count=$(( count + 1 ))
-        if [ "$count" -lt 2 ]; then
-            echo "$count" > "$cf"
-            echo "$(date -Iseconds) $agent: API failure $count of 2, deferring kill" >> "$LOG"
-            return 0
-        fi
-        rm -f "$cf" 2>/dev/null
-    else
-        # CDP port not listening (rc=2): zombie browser, kill immediately as before.
-        rm -f "$STATE_DIR/failcount-$agent" 2>/dev/null
-    fi
-
-    # De-conflict with chromebox-watchdog.sh: if the main browser for this
-    # profile launched <2 min ago, the watchdog just relaunched it — skip the
-    # kill path rather than racing it on a fresh cold start.
-    if recent_relaunch "$cdp_port"; then
-        echo "$(date -Iseconds) $agent: skipping kill (browser launched <2m ago, likely watchdog relaunch)" >> "$LOG"
-        return 0
-    fi
-
-    # Circuit breaker: repeated futile restarts stop here until cooldown.
-    if ! circuit_allows "$agent"; then
-        return 0
-    fi
-
-    restart_browser "$agent" "$cdp_port"
-    # 2026-10-04: post-restart re-check grace extended to ~60s total
-    # (restart_browser sleeps 15s internally + 45s here), matching
-    # chromebox-watchdog.sh's proven 60s retry window. Cold starts on a
-    # loaded box (load ~7) need more than 20s before CDP/API respond.
-    sleep 45
-    if ! check_agent "$agent" "$agent" "$cdp_port"; then
-        echo "$(date -Iseconds) $agent: CRITICAL - still down after restart" >> "$LOG"
-        circuit_note_restart "$agent" fail
-    else
-        echo "$(date -Iseconds) $agent: RECOVERED after restart" >> "$LOG"
-        circuit_note_restart "$agent" ok
-    fi
-}
-
-# Every active node in the NODES.md registry gets checked — new nodes
-# propagate automatically, no per-node blocks to add.
-"$NETVM_BIN/netvm-registry.py" 2>/dev/null | while IFS=: read -r node port; do
-    [ -n "$node" ] && [ -n "$port" ] && check_one "$node" "$port"
-done
-
-# Trim log (keep last 1000 lines)
-tail -1000 "$LOG" > "$LOG.tmp" && mv "$LOG.tmp" "$LOG"

bin/agent-manager.py

@@ -1,905 +0,0 @@
-#!/usr/bin/env python3
-"""agent-manager.py — Agent runs across tailnet devices (stdlib curses).
-
-Read-only dashboard that aggregates agent harness runs (any CLI / bin
-runtime: muse, agy, and whatever else matches the open taxonomy) over
-SSH to every reachable tailnet device, then sorts by agent type.
-
-Surfaces 3 read-only views (no actions in v1):
-  [1] RUNS    — every run, sorted by agent type, then device.
-  [2] TYPES   — counts per agent type with per-device breakdown.
-  [3] DEVICES — per-device reachability, run counts, and notes.
-
-Structure mirrors bin/box-fleet-tui.py: all data-gathering lives in
-pure, testable functions taking injected runners (see gather_*); the
-curses UI is a thin renderer over those functions. The backend differs:
-instead of local repo files, each refresh fans out over tailnet SSH
-(one call per device: tmux panes + process table, joined locally).
-Unreachable devices yield "n/a" rows — never a crash.
-
-Usage:
-  python3 bin/agent-manager.py [--once [--json]]   # non-interactive dump
-"""
-
-from __future__ import annotations
-
-import concurrent.futures
-import curses
-import json
-import os
-import re
-import subprocess
-import sys
-import time
-from datetime import datetime, timezone
-from pathlib import Path
-from typing import Any, Callable, Dict, Iterable, List, Optional, Tuple
-
-REPO_ROOT = Path(__file__).resolve().parent.parent
-
-NA = "n/a"
-
-RunFn = Callable[..., Tuple[int, str]]
-
-# Per-device SSH budget; whole-fleet refresh runs devices in parallel.
-DEVICE_TIMEOUT_S = 15
-SSH_OPTS = ["-o", "BatchMode=yes", "-o", "ConnectTimeout=5"]
-
-# Only these OS classes get an SSH probe (from `tailscale status`).
-SSH_OS = ("linux", "macos")
-
-# =====================================================================
-# Agent-type taxonomy (open: unknown bin runtimes still show up)
-# =====================================================================
-#
-# Classification keys off the argv[0] basename so a wrapper path never
-# matters (/home/super/.local/bin/agy.bin -> agy). Scripts (*.py,
-# *.sh) are never harness runtimes. Anything shaped like a bin runtime
-# (*.bin, *-bin-*) that is not otherwise known shows under its own
-# "other:<basename>" type instead of being dropped.
-
-HARNESS_EXACT = {
-    "agy": "agy",
-    "agy.bin": "agy",
-    "muse-code": "muse",
-    "Muse": "muse",
-    "claude": "claude",
-    "codex": "codex",
-    "gemini": "gemini",
-    "aider": "aider",
-    "opencode": "opencode",
-    "crush": "crush",
-    "amp": "amp",
-}
-
-HARNESS_PREFIX = (
-    ("muse-bin", "muse"),
-)
-
-# Bin-shaped names that are infrastructure, not agent harnesses.
-HARNESS_DENY = frozenset({"tmux.bin", "tmux", "ssh.bin"})
-
-SCRIPT_SUFFIXES = (".py", ".pyc", ".sh", ".pl", ".rb", ".js")
-
-
-def classify_harness(argv0: str) -> Optional[Tuple[str, str]]:
-    """Map an argv[0] to (agent_type, bin_name); None when not a harness.
-
-    Known harnesses collapse to a canonical type ("agy"); unknown bin
-    runtimes keep their own "other:<basename>" type so new harnesses
-    appear without a code change.
-    """
-    base = os.path.basename((argv0 or "").strip().strip("'\""))
-    if not base:
-        return None
-    low = base.lower()
-    if low in HARNESS_DENY:
-        return None
-    if low.endswith(SCRIPT_SUFFIXES):
-        return None
-    if low in HARNESS_EXACT:
-        return HARNESS_EXACT[low], base
-    for prefix, typ in HARNESS_PREFIX:
-        if low.startswith(prefix):
-            return typ, base
-    if low.endswith(".bin") or "-bin-" in low or low.startswith("bin-"):
-        return "other:%s" % base, base
-    return None
-
-
-def type_sort_key(agent_type: str) -> Tuple[int, str]:
-    """Known types first (alpha), then other:* (alpha)."""
-    if agent_type.startswith("other:"):
-        return 1, agent_type
-    return 0, agent_type
-
-
-# =====================================================================
-# Default IO primitives (injectable seams for tests)
-# =====================================================================
-
-def _run(cmd: List[str], timeout: int = 15) -> Tuple[int, str]:
-    """Run cmd, capture output. Returns (returncode, combined_output)."""
-    try:
-        r = subprocess.run(cmd, capture_output=True, text=True,
-                           timeout=timeout)
-        return r.returncode, ((r.stdout or "") + (r.stderr or "")).strip()
-    except subprocess.TimeoutExpired:
-        return 124, "timed out after %ds: %s" % (timeout, " ".join(cmd))
-    except OSError as e:
-        return 127, str(e)
-
-
-def _run_ssh(device: str, remote_cmd: str,
-             timeout: int = DEVICE_TIMEOUT_S,
-             run: Optional[RunFn] = None) -> Tuple[int, str]:
-    """Run one remote command over tailnet SSH. Fails closed, never raises."""
-    run = run or _run
-    try:
-        return run(["ssh"] + SSH_OPTS + [device, remote_cmd],
-                   timeout=timeout)
-    except Exception as e:
-        return 127, str(e)
-
-
-# =====================================================================
-# Pure parsers
-# =====================================================================
-
-def parse_tailscale_status(out: str) -> List[Dict[str, Any]]:
-    """Parse `tailscale status` into [{name, ip, os, online, detail}].
-
-    Unparseable lines are skipped; a warning preamble is ignored.
-    """
-    devices: List[Dict[str, Any]] = []
-    for line in (out or "").splitlines():
-        line = line.rstrip()
-        if not line or line.startswith("Warning:"):
-            continue
-        parts = line.split()
-        if len(parts) < 4:
-            continue
-        ip, name, _user, osname = parts[0], parts[1], parts[2], parts[3]
-        if not re.match(r"^\d+\.\d+\.\d+\.\d+$", ip):
-            continue
-        detail = " ".join(parts[4:]) if len(parts) > 4 else ""
-        online = not detail.startswith("offline")
-        devices.append({"name": name, "ip": ip, "os": osname.lower(),
-                        "online": online, "detail": detail or NA})
-    return devices
-
-
-PANE_PREFIX = "PANE|"
-PS_MARKER = "__PS__"
-
-
-def parse_panes(out: str) -> List[Dict[str, Any]]:
-    """Parse tmux pane lines (PANE|sock|id|pid|session|cmd|title).
-
-    Pane ids repeat across servers, so (socket, id) is the identity.
-    Session groups repeat a pane under several sessions; dedupe by
-    (socket, id), joining session names. The legacy 4-field shape
-    (no socket) still parses with sock="".
-    """
-    seen: Dict[Tuple[str, str], Dict[str, Any]] = {}
-    order: List[Tuple[str, str]] = []
-    for line in (out or "").splitlines():
-        if not line.startswith(PANE_PREFIX):
-            continue
-        fields = line[len(PANE_PREFIX):].split("|")
-        if len(fields) >= 6:
-            sock, pane_id, pid_s, session, cmd = (
-                fields[0].strip(), fields[1].strip(), fields[2].strip(),
-                fields[3].strip(), fields[4].strip())
-            title = "|".join(fields[5:]).strip()
-        elif len(fields) >= 4:
-            sock, pane_id, pid_s, session, cmd = (
-                "", fields[0].strip(), fields[1].strip(),
-                fields[2].strip(), fields[3].strip())
-            title = "|".join(fields[4:]).strip()
-        else:
-            continue
-        try:
-            pid = int(pid_s)
-        except ValueError:
-            continue
-        session = session or NA
-        key = (sock, pane_id)
-        if key in seen:
-            prev = seen[key]["session"]
-            if session not in prev.split(","):
-                seen[key]["session"] = prev + "," + session
-            continue
-        seen[key] = {"sock": sock, "pane": pane_id, "pid": pid,
-                     "session": session, "cmd": cmd, "title": title}
-        order.append(key)
-    return [seen[k] for k in order]
-
-
-def display_session(run: Dict[str, Any]) -> str:
-    """Session label; socket-qualified unless it is the default server."""
-    session = str(run.get("session", NA))
-    sock = run.get("sock") or ""
-    if session == NA or sock in ("", "default"):
-        return session
-    return "%s:%s" % (sock, session)
-
-
-def parse_ps(out: str) -> Dict[int, Dict[str, Any]]:
-    """Parse `ps -eo pid,ppid,etime,command` into {pid: rec}."""
-    procs: Dict[int, Dict[str, Any]] = {}
-    for line in (out or "").splitlines():
-        parts = line.split(None, 3)
-        if len(parts) < 4:
-            continue
-        try:
-            pid, ppid = int(parts[0]), int(parts[1])
-        except ValueError:
-            continue  # header row
-        procs[pid] = {"pid": pid, "ppid": ppid, "etime": parts[2],
-                      "args": parts[3]}
-    return procs
-
-
-def split_scan(out: str) -> Tuple[str, str]:
-    """Split a device scan into (pane_text, ps_text) at the marker."""
-    if PS_MARKER in out:
-        pane_text, _, ps_text = out.partition(PS_MARKER)
-        return pane_text, ps_text
-    return out, ""
-
-
-def _descendants(procs: Dict[int, Dict[str, Any]],
-                 root: int) -> List[int]:
-    """Pids under root (breadth-first via ppid links)."""
-    kids: Dict[int, List[int]] = {}
-    for pid, rec in procs.items():
-        kids.setdefault(rec["ppid"], []).append(pid)
-    out: List[int] = []
-    queue = list(kids.get(root, []))
-    seen = {root}
-    while queue:
-        pid = queue.pop(0)
-        if pid in seen:
-            continue
-        seen.add(pid)
-        out.append(pid)
-        queue.extend(kids.get(pid, []))
-    return out
-
-
-def join_runs(panes: List[Dict[str, Any]],
-              procs: Dict[int, Dict[str, Any]]) -> List[Dict[str, Any]]:
-    """Join tmux panes with harness processes into run records.
-
-    A pane is a run when its current command classifies as a harness
-    or a harness binary runs among its descendants. Harness processes
-    under no pane surface as bare runs (session/pane n/a). Returns
-    records sorted by (agent_type, session, pane).
-    """
-    runs: List[Dict[str, Any]] = []
-    claimed: set = set()
-    pane_roots = {p["pid"] for p in panes}
-
-    for pane in panes:
-        argv = (pane.get("cmd") or "").strip()
-        pane_hit = classify_harness(argv.split()[0] if argv else "")
-        # Always resolve the live harness descendant: the pane's root
-        # is usually a shell, so its pid/etime/bin would mislead.
-        # The pane-command match is only a fallback (stale command).
-        hit = None
-        hpid: Optional[int] = None
-        for pid in _descendants(procs, pane["pid"]):
-            rec = procs.get(pid)
-            if not rec:
-                continue
-            first = rec["args"].split()[0] if rec["args"] else ""
-            hit = classify_harness(first)
-            if hit is not None:
-                hpid = pid
-                break
-        if hit is None:
-            if pane_hit is None:
-                continue
-            hit, hpid = pane_hit, pane["pid"]
-        agent_type, _bin = hit
-        claimed.add(hpid)
-        prec = procs.get(hpid, {})
-        if prec.get("args"):
-            binn = os.path.basename(prec["args"].split()[0])
-        elif argv:
-            binn = argv.split()[0]
-        else:
-            binn = NA
-        runs.append({
-            "type": agent_type,
-            "bin": binn,
-            "sock": pane.get("sock", ""),
-            "session": pane.get("session", NA),
-            "pane": pane.get("pane", NA),
-            "pid": hpid,
-            "etime": prec.get("etime", NA),
-            "title": (pane.get("title") or "")[:48],
-        })
-
-    # Bare harness processes (no tmux pane above them).
-    for pid, rec in procs.items():
-        if pid in claimed:
-            continue
-        first = rec["args"].split()[0] if rec["args"] else ""
-        hit = classify_harness(first)
-        if hit is None:
-            continue
-        # Skip when some pane root is an ancestor (already covered).
-        anc, under_pane = rec["ppid"], False
-        hops = 0
-        while anc in procs and hops < 64:
-            if anc in pane_roots:
-                under_pane = True
-                break
-            anc = procs[anc]["ppid"]
-            hops += 1
-        if under_pane:
-            continue
-        claimed.add(pid)
-        runs.append({
-            "type": hit[0],
-            "bin": os.path.basename(first),
-            "sock": "",
-            "session": NA,
-            "pane": NA,
-            "pid": pid,
-            "etime": rec.get("etime", NA),
-            "title": "",
-        })
-
-    runs.sort(key=lambda r: (type_sort_key(r["type"]),
-                             str(r["session"]), str(r["pane"])))
-    return runs
-
-
-# =====================================================================
-# Surfaces: devices + runs
-# =====================================================================
-
-PANE_FORMAT = ("PANE|%s|#{pane_id}|#{pane_pid}|#{session_name}|"
-               "#{pane_current_command}|#{pane_title}")
-
-REMOTE_SCAN_CMD = (
-    "for d in \"${TMUX_TMPDIR:-/tmp}/tmux-$(id -u)\" "
-    "\"${TMPDIR:-/tmp}/tmux-$(id -u)\" /tmp/tmux-$(id -u); do "
-    "for s in \"$d\"/*; do [ -S \"$s\" ] || continue; "
-    "n=$(basename \"$s\"); "
-    "tmux -S \"$s\" list-panes -a -F \"PANE|$n|#{pane_id}|#{pane_pid}|"
-    "#{session_name}|#{pane_current_command}|#{pane_title}\" "
-    "2>/dev/null; done; done; "
-    "echo '%s'; ps -eo pid,ppid,etime,command 2>/dev/null" % PS_MARKER
-)
-
-
-def local_tmux_sockets() -> List[str]:
-    """Absolute tmux socket paths on this machine (may be empty)."""
-    uid = os.getuid() if hasattr(os, "getuid") else 0
-    cands = []
-    for base in (os.environ.get("TMUX_TMPDIR") or "/tmp",
-                 os.environ.get("TMPDIR") or "/tmp", "/tmp"):
-        cands.append(os.path.join(base, "tmux-%d" % uid))
-    found: List[str] = []
-    seen_dirs = set()
-    for d in cands:
-        if d in seen_dirs:
-            continue
-        seen_dirs.add(d)
-        try:
-            names = sorted(os.listdir(d))
-        except Exception:
-            continue
-        for n in names:
-            p = os.path.join(d, n)
-            try:
-                import stat
-                if stat.S_ISSOCK(os.stat(p).st_mode):
-                    found.append(p)
-            except Exception:
-                continue
-    # Same server via two spellings: keep first per basename.
-    dedup: List[str] = []
-    seen_base = set()
-    for p in found:
-        b = os.path.basename(p)
-        if b not in seen_base:
-            seen_base.add(b)
-            dedup.append(p)
-    return dedup
-
-
-def local_device_name() -> str:
-    """Short hostname of this machine (never raises)."""
-    try:
-        import socket
-        return socket.gethostname().split(".")[0]
-    except Exception:
-        return "localhost"
-
-
-def gather_devices(run: Optional[RunFn] = None,
-                   local_name: Optional[str] = None) -> Dict[str, Any]:
-    """Tailnet devices from `tailscale status` + the local machine.
-
-    Returns {"devices": [{name, ip, os, online, local, ssh, detail}],
-    "note": str}. Devices are stable-sorted: local first, then by name.
-    "ssh" marks whether v1 probes the device (online + ssh-capable OS).
-    """
-    run = run or _run
-    local_name = local_name or local_device_name()
-    rc, out = run(["tailscale", "status"], timeout=10)
-    if rc != 0:
-        return {"devices": [{"name": local_name, "ip": NA, "os": NA,
-                             "online": True, "local": True, "ssh": False,
-                             "detail": "local only"}],
-                "note": "tailscale status failed (%s); local only."
-                % (out.strip().splitlines()[-1][:80] if out.strip()
-                   else "rc=%d" % rc)}
-    devices = []
-    for d in parse_tailscale_status(out):
-        local = (d["name"] == local_name)
-        ssh = bool(d["online"] and d["os"] in SSH_OS and not local)
-        devices.append({"name": d["name"], "ip": d["ip"], "os": d["os"],
-                        "online": d["online"], "local": local, "ssh": ssh,
-                        "detail": d["detail"]})
-    if not any(d["local"] for d in devices):
-        devices.append({"name": local_name, "ip": NA, "os": NA,
-                        "online": True, "local": True, "ssh": False,
-                        "detail": "local"})
-    devices.sort(key=lambda d: (not d["local"], d["name"]))
-    return {"devices": devices, "note": ""}
-
-
-def gather_device_runs(device: Dict[str, Any],
-                       run: Optional[RunFn] = None) -> Dict[str, Any]:
-    """One device scan -> {device, ok, runs, note}. Never raises."""
-    run = run or _run
-    name = device.get("name", "?")
-    try:
-        if device.get("local"):
-            pane_chunks = []
-            for sock in local_tmux_sockets():
-                rc1, chunk = run(
-                    ["tmux", "-S", sock, "list-panes", "-a", "-F",
-                     PANE_FORMAT % os.path.basename(sock)],
-                    timeout=DEVICE_TIMEOUT_S)
-                if rc1 == 0 and chunk:
-                    pane_chunks.append(chunk)
-            _rc2, ps_out = run(["ps", "-eo", "pid,ppid,etime,command"],
-                               timeout=DEVICE_TIMEOUT_S)
-            out = ("\n".join(pane_chunks) + "\n" + PS_MARKER + "\n"
-                   + ps_out)
-        else:
-            rc, out = _run_ssh(name, REMOTE_SCAN_CMD, run=run)
-            ssh_err = "" if rc == 0 else out.strip().splitlines()
-            ssh_err = ssh_err[-1][:100] if ssh_err else "rc=%d" % rc
-            if rc != 0:
-                return {"device": name, "ok": False, "runs": [],
-                        "note": "ssh failed: %s" % ssh_err}
-        pane_text, ps_text = split_scan(out)
-        runs = join_runs(parse_panes(pane_text), parse_ps(ps_text))
-        for r in runs:
-            r["device"] = name
-        return {"device": name, "ok": True, "runs": runs, "note": ""}
-    except Exception as e:
-        return {"device": name, "ok": False, "runs": [],
-                "note": "scan failed: %s" % e}
-
-
-def gather_all(run: Optional[RunFn] = None,
-               devices: Optional[List[Dict[str, Any]]] = None,
-               max_workers: int = 8) -> Dict[str, Any]:
-    """One-shot snapshot: devices + runs sorted by agent type.
-
-    Devices scan in parallel (threads); each device is isolated — one
-    failure never blocks the rest. Returns {"devices": [...],
-    "runs": [...] (sorted by type/device), "by_type": {type: {total,
-    devices: {name: n}}}, "unreachable": [names], "note": str}.
-    """
-    run = run or _run
-    dev_info = gather_devices(run=run)
-    if devices is None:
-        devices = [d for d in dev_info["devices"]
-                   if d.get("local") or d.get("ssh")]
-    else:
-        devices = [d for d in devices
-                   if d.get("local") or d.get("ssh")]
-    probed = {d["name"] for d in devices}
-
-    results: List[Dict[str, Any]] = []
-    if devices:
-        with concurrent.futures.ThreadPoolExecutor(
-                max_workers=min(max_workers, len(devices))) as pool:
-            futs = {pool.submit(gather_device_runs, d, run): d["name"]
-                    for d in devices}
-            for fut in concurrent.futures.as_completed(futs):
-                try:
-                    results.append(fut.result())
-                except Exception as e:
-                    results.append({"device": futs[fut], "ok": False,
-                                    "runs": [], "note": "scan error: %s" % e})
-
-    runs: List[Dict[str, Any]] = []
-    unreachable: List[str] = []
-    for res in results:
-        if not res.get("ok"):
-            unreachable.append(res["device"])
-            continue
-        runs.extend(res.get("runs", []))
-    runs.sort(key=lambda r: (type_sort_key(r["type"]), r.get("device", ""),
-                             str(r.get("session", ""))))
-
-    by_type: Dict[str, Dict[str, Any]] = {}
-    for r in runs:
-        bucket = by_type.setdefault(r["type"], {"total": 0, "devices": {}})
-        bucket["total"] += 1
-        dev = r.get("device", "?")
-        bucket["devices"][dev] = bucket["devices"].get(dev, 0) + 1
-
-    skipped = sorted(d["name"] for d in dev_info["devices"]
-                     if d["name"] not in probed)
-    notes = [dev_info["note"]] if dev_info["note"] else []
-    if skipped:
-        notes.append("skipped (offline/mobile/key): %s" % ", ".join(skipped))
-    return {"devices": dev_info["devices"], "runs": runs,
-            "by_type": by_type, "unreachable": sorted(unreachable),
-            "note": "; ".join(notes)}
-
-
-# =====================================================================
-# Curses UI (thin read-only renderer over gather_*)
-# =====================================================================
-
-AUTO_REFRESH_S = 30.0
-
-
-class AgentManagerTUI:
-    """Read-only agent-run console. q quits, r refreshes, ? helps."""
-
-    def __init__(self, stdscr: "curses.window"):
-        self.stdscr = stdscr
-        self.current_tab = 0
-        self.tabs = [
-            "1: RUNS",
-            "2: TYPES",
-            "3: DEVICES",
-        ]
-        try:
-            curses.curs_set(0)
-        except Exception:
-            pass
-        self.stdscr.nodelay(True)
-        self.stdscr.keypad(True)
-        if hasattr(curses, "set_escdelay"):
-            try:
-                curses.set_escdelay(25)
-            except Exception:
-                pass
-        self._init_colors()
-
-        self.scroll = 0
-        self.show_help = False
-        self.status_msg = "Scanning tailnet devices..."
-        self.last_refresh = 0.0
-        self.snapshot: Dict[str, Any] = {}
-        self.refresh()
-
-    # -- setup ------------------------------------------------------
-
-    def _init_colors(self) -> None:
-        try:
-            curses.start_color()
-            curses.use_default_colors()
-            curses.init_pair(1, curses.COLOR_CYAN, -1)
-            curses.init_pair(2, curses.COLOR_YELLOW, -1)
-            curses.init_pair(3, curses.COLOR_GREEN, -1)
-            curses.init_pair(4, curses.COLOR_RED, -1)
-            curses.init_pair(5, curses.COLOR_MAGENTA, -1)
-            curses.init_pair(6, curses.COLOR_BLACK, curses.COLOR_CYAN)
-            curses.init_pair(7, curses.COLOR_BLACK, curses.COLOR_WHITE)
-            curses.init_pair(8, curses.COLOR_BLACK, curses.COLOR_YELLOW)
-        except Exception:
-            pass
-
-    def _attr(self, name: str) -> int:
-        try:
-            mapping = {
-                "normal": curses.color_pair(0),
-                "cyan": curses.color_pair(1) | curses.A_BOLD,
-                "yellow": curses.color_pair(2) | curses.A_BOLD,
-                "green": curses.color_pair(3) | curses.A_BOLD,
-                "red": curses.color_pair(4) | curses.A_BOLD,
-                "magenta": curses.color_pair(5) | curses.A_BOLD,
-                "head_sel": curses.color_pair(6) | curses.A_BOLD,
-                "row_sel": curses.color_pair(7) | curses.A_BOLD,
-                "warn": curses.color_pair(8) | curses.A_BOLD,
-                "dim": curses.A_DIM,
-            }
-            return mapping.get(name, 0)
-        except Exception:
-            return 0
-
-    # -- data -------------------------------------------------------
-
-    def refresh(self) -> None:
-        try:
-            self.snapshot = gather_all()
-            runs = len(self.snapshot.get("runs", []))
-            devs = len([d for d in self.snapshot.get("devices", [])
-                        if d.get("local") or d.get("ssh")])
-            self.status_msg = (
-                "Snapshot %s: %d runs on %d devices "
-                "(auto-refresh %ds; r=refresh)" % (
-                    datetime.now().strftime("%H:%M:%S"), runs, devs,
-                    int(AUTO_REFRESH_S)))
-        except Exception as e:
-            self.snapshot = {}
-            self.status_msg = "Refresh failed (showing n/a): %s" % e
-        self.last_refresh = time.time()
-        self.scroll = 0
-
-    # -- render helpers ---------------------------------------------
-
-    def safe_addstr(self, y: int, x: int, text: str, attr: int = 0) -> None:
-        h, w = self.stdscr.getmaxyx()
-        if 0 <= y < h and 0 <= x < w:
-            try:
-                self.stdscr.addstr(y, x, text[:max(0, w - x - 1)], attr)
-            except Exception:
-                pass
-
-    def _render_header(self, w: int) -> None:
-        self.safe_addstr(0, 0, " " * w, self._attr("head_sel"))
-        title = " AGENT MANAGER (read-only)  [agent-manager.py] "
-        self.safe_addstr(0, 1, title, self._attr("head_sel"))
-        self.safe_addstr(1, 0, " " * w, self._attr("dim"))
-        col = 1
-        for idx, tab_name in enumerate(self.tabs):
-            pill = " [%s] " % tab_name
-            attr = self._attr("head_sel") if idx == self.current_tab \
-                else self._attr("dim")
-            self.safe_addstr(1, col, pill, attr)
-            col += len(pill) + 1
-        self.safe_addstr(2, 0, "-" * w, self._attr("dim"))
-
-    def _render_footer(self, h: int, w: int) -> None:
-        self.safe_addstr(h - 2, 0, "-" * w, self._attr("dim"))
-        hints = " 1-3/Tab: Tabs  j/k: Scroll  r: Refresh  ?: Help  q: Quit"
-        self.safe_addstr(h - 1, 1, self.status_msg[: w - 2],
-                         self._attr("dim"))
-        if len(self.status_msg) + len(hints) + 2 < w:
-            self.safe_addstr(h - 1, w - len(hints) - 1, hints,
-                             self._attr("dim"))
-
-    def _body(self, h: int, w: int, title: str,
-              lines: List[Tuple[str, str]]) -> None:
-        self.safe_addstr(3, 2, title, self._attr("cyan"))
-        self.safe_addstr(4, 2, "-" * (w - 4), self._attr("dim"))
-        max_rows = h - 8
-        visible = lines[self.scroll:self.scroll + max_rows]
-        for i, (text, attr_name) in enumerate(visible):
-            self.safe_addstr(5 + i, 2, text, self._attr(attr_name))
-        if self.scroll > 0:
-            self.safe_addstr(5, w - 6, "^more", self._attr("dim"))
-        if self.scroll + max_rows < len(lines):
-            self.safe_addstr(h - 3, w - 6, "vmore", self._attr("dim"))
-
-    def _note_lines(self, w: int) -> List[Tuple[str, str]]:
-        note = self.snapshot.get("note", "")
-        unreach = self.snapshot.get("unreachable", [])
-        lines: List[Tuple[str, str]] = []
-        if unreach:
-            lines.append(("", "normal"))
-            lines.append(("unreachable: %s" % ", ".join(unreach),
-                          "red"))
-        if note:
-            lines.append(("", "normal"))
-            lines.append(("note: %s" % note[: w - 10], "yellow"))
-        return lines
-
-    # -- per-tab renderers ------------------------------------------
-
-    def _render_runs(self, h: int, w: int) -> None:
-        runs = self.snapshot.get("runs", [])
-        lines: List[Tuple[str, str]] = [
-            ("%-14s %-12s %-16s %-14s %-6s %-11s %s"
-             % ("TYPE", "DEVICE", "BIN", "SESSION", "PANE", "ELAPSED",
-                "TITLE"), "dim"),
-        ]
-        last_type = None
-        for r in runs:
-            typ = r.get("type", "?")
-            if typ != last_type:
-                lines.append(("", "normal"))
-                last_type = typ
-            attr = "green" if not typ.startswith("other:") else "yellow"
-            lines.append((
-                "%-14s %-12s %-16s %-14s %-6s %-11s %s" % (
-                    typ[:14], r.get("device", "?")[:12],
-                    r.get("bin", NA)[:16], display_session(r)[:14],
-                    str(r.get("pane", NA))[:6],
-                    str(r.get("etime", NA))[:11],
-                    r.get("title", "")[: w - 80]), attr))
-        if not runs:
-            lines.append(("(No agent runs found on probed devices.)",
-                          "dim"))
-        lines.extend(self._note_lines(w))
-        self._body(h, w, "AGENT RUNS SORTED BY TYPE (%d)" % len(runs),
-                   lines)
-
-    def _render_types(self, h: int, w: int) -> None:
-        by_type = self.snapshot.get("by_type", {})
-        lines: List[Tuple[str, str]] = []
-        total = sum(b.get("total", 0) for b in by_type.values())
-        lines.append(("Agent types: %d  |  total runs: %d"
-                      % (len(by_type), total), "cyan"))
-        lines.append(("", "normal"))
-        for typ in sorted(by_type, key=type_sort_key):
-            bucket = by_type[typ]
-            attr = "green" if not typ.startswith("other:") else "yellow"
-            lines.append(("%-16s %d" % (typ, bucket.get("total", 0)),
-                          attr))
-            for dev, n in sorted(bucket.get("devices", {}).items()):
-                lines.append(("  %-14s %d" % (dev, n), "normal"))
-            lines.append(("", "normal"))
-        if not by_type:
-            lines.append(("(No agent types observed.)", "dim"))
-        lines.extend(self._note_lines(w))
-        self._body(h, w, "COUNTS BY AGENT TYPE", lines)
-
-    def _render_devices(self, h: int, w: int) -> None:
-        devices = self.snapshot.get("devices", [])
-        runs = self.snapshot.get("runs", [])
-        counts: Dict[str, int] = {}
-        for r in runs:
-            dev = r.get("device", "?")
-            counts[dev] = counts.get(dev, 0) + 1
-        unreach = set(self.snapshot.get("unreachable", []))
-        lines: List[Tuple[str, str]] = [
-            ("%-24s %-15s %-7s %-7s %-5s %s"
-             % ("DEVICE", "IP", "OS", "PROBED", "RUNS", "DETAIL"), "dim"),
-        ]
-        for d in devices:
-            name = d.get("name", "?")
-            probed = bool(d.get("local") or d.get("ssh"))
-            if name in unreach:
-                attr = "red"
-            elif not probed:
-                attr = "dim"
-            elif counts.get(name):
-                attr = "green"
-            else:
-                attr = "normal"
-            lines.append((
-                "%-24s %-15s %-7s %-7s %-5s %s" % (
-                    name[:24] + (" *" if d.get("local") else ""),
-                    d.get("ip", NA)[:15], str(d.get("os", NA))[:7],
-                    "yes" if probed else "no",
-                    counts.get(name, 0) if probed else NA,
-                    str(d.get("detail", ""))[: w - 66]), attr))
-        lines.append(("", "normal"))
-        lines.append(("* = local (no SSH); unreachable shows n/a, never "
-                      "blocks the rest.", "dim"))
-        lines.extend(self._note_lines(w))
-        self._body(h, w, "TAILNET DEVICES", lines)
-
-    def _render_help(self, h: int, w: int) -> None:
-        modal_w = min(64, w - 6)
-        modal_h = 13
-        top = (h - modal_h) // 2
-        left = (w - modal_w) // 2
-        for y in range(top, top + modal_h):
-            self.safe_addstr(y, left, " " * modal_w, self._attr("row_sel"))
-        self.safe_addstr(top, left, "+" + "-" * (modal_w - 2) + "+",
-                         self._attr("cyan"))
-        for y in range(top + 1, top + modal_h - 1):
-            self.safe_addstr(y, left, "|", self._attr("cyan"))
-            self.safe_addstr(y, left + modal_w - 1, "|",
-                             self._attr("cyan"))
-        self.safe_addstr(top + modal_h - 1, left,
-                         "+" + "-" * (modal_w - 2) + "+",
-                         self._attr("cyan"))
-        self.safe_addstr(top + 1, left + 3, "AGENT MANAGER HELP (read-only)",
-                         self._attr("cyan"))
-        for i, hint in enumerate([
-            "1-3 / Tab: switch surfaces",
-            "j/k / Up/Down: scroll",
-            "r: refresh snapshot now",
-            "q / Esc: quit (Esc closes help first)",
-            "",
-            "One SSH call per device per refresh.",
-            "Missing data shows as 'n/a' — never a crash.",
-        ]):
-            self.safe_addstr(top + 3 + i, left + 4, hint,
-                             self._attr("normal"))
-
-    # -- input + main loop ------------------------------------------
-
-    def _handle_key(self, ch: int) -> bool:
-        if ch in (3, 4):  # Ctrl+C / Ctrl+D
-            return False
-        if self.show_help:
-            if ch in (27, ord("q"), ord("Q"), ord("?")):
-                self.show_help = False
-            return True
-        if ch in (ord("q"), ord("Q")):
-            return False
-        if ch in (27, ord("?")):
-            self.show_help = True
-            return True
-        if ch in (ord("1"), ord("2"), ord("3")):
-            self.current_tab = ch - ord("1")
-            self.scroll = 0
-            return True
-        if ch == ord("\t"):
-            self.current_tab = (self.current_tab + 1) % len(self.tabs)
-            self.scroll = 0
-            return True
-        if ch in (ord("j"), curses.KEY_DOWN):
-            self.scroll += 1
-            return True
-        if ch in (ord("k"), curses.KEY_UP):
-            self.scroll = max(0, self.scroll - 1)
-            return True
-        if ch in (ord("r"), ord("R")):
-            self.refresh()
-            return True
-        return True
-
-    def run(self) -> None:
-        renderers = [self._render_runs, self._render_types,
-                     self._render_devices]
-        while True:
-            h, w = self.stdscr.getmaxyx()
-            self.stdscr.erase()
-            self._render_header(w)
-            try:
-                renderers[self.current_tab](h, w)
-            except Exception as e:
-                self.safe_addstr(5, 4, "Render error (n/a): %s" % e,
-                                 self._attr("red"))
-            self._render_footer(h, w)
-            if self.show_help:
-                self._render_help(h, w)
-            self.stdscr.refresh()
-            try:
-                ch = self.stdscr.getch()
-                if ch != -1 and not self._handle_key(ch):
-                    break
-            except KeyboardInterrupt:
-                break
-            if time.time() - self.last_refresh > AUTO_REFRESH_S:
-                self.refresh()
-            time.sleep(0.05)
-
-
-def main(argv: Optional[List[str]] = None) -> int:
-    argv = list(sys.argv[1:] if argv is None else argv)
-    if "--once" in argv:
-        snap = gather_all()
-        if "--json" in argv:
-            print(json.dumps(snap, indent=2, default=str))
-        else:
-            print("== RUNS (%d) ==" % len(snap.get("runs", [])))
-            for r in snap.get("runs", []):
-                print("%-14s %-12s %-16s %s/%s %s" % (
-                    r.get("type"), r.get("device"), r.get("bin"),
-                    display_session(r), r.get("pane"), r.get("etime")))
-            print("== BY TYPE == ")
-            for typ, b in sorted(snap.get("by_type", {}).items()):
-                print("%s: %d %s" % (typ, b["total"], b["devices"]))
-            print("unreachable: %s" % snap.get("unreachable"))
-            if snap.get("note"):
-                print("note: %s" % snap["note"])
-        return 0
-    curses.wrapper(lambda stdscr: AgentManagerTUI(stdscr).run())
-    return 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/agent_cognitive_probe.py

@@ -0,0 +1 @@
+agent-cognitive-probe.py

bin/agent_md.py

@@ -1,625 +0,0 @@
-#!/usr/bin/env python3
-"""agent_md.py — Access and modify Muse agent .md files via Hatch gateway and SSH.
-
-Enables operators to inspect, audit, diff, and inject operational DRIVE into
-Muse agents across the fleet (muse, pip, 646, opm, def, dev).
-"""
-
-import difflib
-import json
-import os
-import re
-import sys
-import time
-from pathlib import Path
-
-# Ensure muse_cli can be imported
-sys.path.insert(0, os.path.expanduser("~/.local/lib/python3.14/site-packages"))
-try:
-    from muse_cli.gateway import Gateway, load_cookies, AuthError, GatewayError
-except ImportError:
-    Gateway = None
-
-NETVM_ROOT = Path("/home/super/Projects/NetVM")
-SHARED_OPERATORS = NETVM_ROOT / "shared" / "operators"
-VALID_ACCOUNTS = ["muse", "pip", "646", "opm", "def", "dev"]
-
-TARGET_MD_FILES = [
-    "SOUL.md",
-    "PROACTIVE_PREFERENCES.md",
-    "HEARTBEAT.md",
-    "AGENTS.md",
-    "MEMORY.md",
-    "USER.md",
-    "TOOLS.md",
-    "IDENTITY.md",
-]
-
-
-class MDValidationError(ValueError):
-    """An md account/filename/path failed safety validation.
-
-    box-ctl.py maps this to BAD_NAME; it is always raised before any
-    gateway call or filesystem write.
-    """
-
-
-MD_ACCOUNT_RE = re.compile(r"^[A-Za-z0-9][A-Za-z0-9_-]{0,31}$")
-MD_FILENAME_RE = re.compile(r"^[A-Za-z0-9_.-]{1,128}$")
-MD_SUBPATH_RE = re.compile(r"^[A-Za-z0-9_.-]+(/[A-Za-z0-9_.-]+)*$")
-
-
-def validate_account(account: str) -> str:
-    """Reject account values that could escape the cookies/config path."""
-    if not isinstance(account, str) or not MD_ACCOUNT_RE.fullmatch(account):
-        raise MDValidationError(
-            "Invalid agent account %r: must match ^[A-Za-z0-9][A-Za-z0-9_-]{0,31}$"
-            % (account,))
-    return account
-
-
-def validate_filename(filename: str, template_only: bool = False) -> str:
-    """Reject filenames that could escape the md directory.
-
-    Template flows (diff/amend/append/pull) additionally require one of
-    TARGET_MD_FILES, since they index into shared/operators/.
-    """
-    if template_only:
-        if filename not in TARGET_MD_FILES:
-            raise MDValidationError(
-                "Unknown shared template %r: must be one of %s"
-                % (filename, sorted(TARGET_MD_FILES)))
-        return filename
-    if not isinstance(filename, str) or filename in (".", "..") \
-            or not MD_FILENAME_RE.fullmatch(filename):
-        raise MDValidationError(
-            "Invalid filename %r: plain basename, no directories" % (filename,))
-    return filename
-
-
-def validate_subpath(path: str) -> str:
-    """Reject list paths that escape the container root ('' = root)."""
-    if path in (None, ""):
-        return ""
-    if not isinstance(path, str) or not MD_SUBPATH_RE.fullmatch(path) \
-            or ".." in path.split("/"):
-        raise MDValidationError(
-            "Invalid list path %r: subdir without '..'" % (path,))
-    return path
-
-# Tunnel / Port inventory
-TUNNEL_PORTS = {
-    "muse-main": {"port": 2224, "terminal": 7681, "user": "muse"},
-    "muse":      {"port": 2225, "terminal": 7682, "user": "hatch"},
-    "646":       {"port": 2226, "terminal": 7683, "user": "hatch"},
-    "pip":       {"port": 2227, "terminal": 7684, "user": "hatch"},
-    "opm":       {"port": 2228, "terminal": 7685, "user": "hatch"},
-    "def":       {"port": 2229, "terminal": 7686, "user": "hatch"},
-    "dev":       {"port": 2230, "terminal": 7687, "user": "hatch"},
-}
-
-
-def get_gateway(account: str) -> "Gateway":
-    """Obtain an authenticated Gateway connection for an account."""
-    validate_account(account)
-    if not Gateway:
-        raise RuntimeError("muse_cli.gateway module is not available")
-    conf_dir = Path.home() / ".config" / "muse-cli" / account
-    cfile = conf_dir / "cookies.txt"
-    if not cfile.exists():
-        raise FileNotFoundError(f"No cookies found for account '{account}' at {cfile}")
-    cookies = load_cookies(str(cfile))
-    if not cookies.strip():
-        raise ValueError(f"Cookies file for '{account}' is empty")
-    return Gateway(cookies)
-
-
-def list_files(account: str, path: str = "") -> list:
-    """List files in the agent container filesystem via Hatch."""
-    path = validate_subpath(path)
-    gw = get_gateway(account)
-    res = gw.call_json("fs.list", body={"path": path})
-    return res.get("entries", [])
-
-
-def read_md(account: str, filename: str, max_bytes: int = 200000) -> dict:
-    """Read a markdown file from the agent container via Hatch."""
-    validate_filename(filename)
-    gw = get_gateway(account)
-    offset = 0
-    chunks = []
-    chunk_len = min(65536, max_bytes)
-    
-    while True:
-        body = {"path": filename, "offset": offset, "len": chunk_len}
-        res = gw.call_json("fs.read", body=body)
-        text = res.get("text", "")
-        if not text and res.get("data_base64"):
-            import base64
-            text = base64.b64decode(res["data_base64"]).decode("utf-8", "replace")
-        chunks.append(text)
-        offset += len(text.encode("utf-8"))
-        if res.get("eof") or offset >= max_bytes or not text:
-            break
-            
-    full_text = "".join(chunks)
-    return {
-        "ok": True,
-        "account": account,
-        "filename": filename,
-        "size": len(full_text.encode("utf-8")),
-        "text": full_text,
-        "eof": True,
-    }
-
-
-def write_md(account: str, filename: str, text: str, overwrite: bool = True, append: bool = False) -> dict:
-    """Write content to a file in the agent container via Hatch."""
-    validate_filename(filename)
-    gw = get_gateway(account)
-    body = {
-        "path": filename,
-        "overwrite": overwrite,
-        "append": append,
-        "create_parent": True,
-        "text": text,
-    }
-    res = gw.call_json("fs.write", body=body)
-    return {
-        "ok": True,
-        "account": account,
-        "filename": filename,
-        "bytes_written": res.get("bytes_written", len(text.encode("utf-8"))),
-        "path": res.get("path", f"/{filename}"),
-    }
-
-
-def audit_agents(accounts: list = None) -> dict:
-    """Audit markdown files and operational DRIVE across fleet agents."""
-    accounts = accounts or VALID_ACCOUNTS
-    for acct in accounts:
-        validate_account(acct)
-    results = {}
-    
-    for acct in accounts:
-        acct_res = {
-            "account": acct,
-            "connected": False,
-            "files": {},
-            "drive_status": {},
-            "issues": [],
-            "drive_score": 0,
-        }
-        try:
-            gw = get_gateway(acct)
-            acct_res["connected"] = True
-            acct_res["vm_id"] = gw.vm_id
-            
-            entries = gw.call_json("fs.list", body={"path": ""}).get("entries", [])
-            entry_map = {e["name"]: e for e in entries}
-            
-            for tf in TARGET_MD_FILES:
-                if tf in entry_map:
-                    info = entry_map[tf]
-                    acct_res["files"][tf] = {
-                        "exists": True,
-                        "size": info.get("size", 0),
-                        "modified": info.get("modifiedAt", ""),
-                    }
-                else:
-                    acct_res["files"][tf] = {
-                        "exists": False,
-                        "size": 0,
-                        "modified": None,
-                    }
-
-            # Analyze DRIVE indicators
-            # 1. HEARTBEAT.md checklist
-            hb_info = acct_res["files"].get("HEARTBEAT.md", {})
-            if not hb_info.get("exists"):
-                acct_res["drive_status"]["heartbeat"] = "MISSING"
-                acct_res["issues"].append("HEARTBEAT.md missing (no recurring checks)")
-            elif hb_info.get("size", 0) <= 120:
-                acct_res["drive_status"]["heartbeat"] = "EMPTY_CHECKLIST"
-                acct_res["issues"].append("HEARTBEAT.md has empty checklist (background runner idle)")
-            else:
-                acct_res["drive_status"]["heartbeat"] = "ACTIVE"
-                acct_res["drive_score"] += 25
-
-            # 2. PROACTIVE_PREFERENCES.md
-            pro_info = acct_res["files"].get("PROACTIVE_PREFERENCES.md", {})
-            if not pro_info.get("exists"):
-                acct_res["drive_status"]["proactive"] = "MISSING"
-                acct_res["issues"].append("PROACTIVE_PREFERENCES.md missing")
-            elif pro_info.get("size", 0) <= 500:
-                acct_res["drive_status"]["proactive"] = "BLANK_TEMPLATE"
-                acct_res["issues"].append("PROACTIVE_PREFERENCES.md unconfigured (never reaches out)")
-            else:
-                acct_res["drive_status"]["proactive"] = "CONFIGURED"
-                acct_res["drive_score"] += 25
-
-            # 3. SOUL.md
-            soul_info = acct_res["files"].get("SOUL.md", {})
-            if not soul_info.get("exists"):
-                acct_res["drive_status"]["soul"] = "MISSING"
-                acct_res["issues"].append("SOUL.md missing")
-            elif soul_info.get("size", 0) <= 850:
-                acct_res["drive_status"]["soul"] = "PASSIVE_STOCK"
-                acct_res["issues"].append("SOUL.md is passive stock template (no operator drive)")
-            else:
-                acct_res["drive_status"]["soul"] = "OPERATOR_SOUL"
-                acct_res["drive_score"] += 25
-
-            # 4. TOOLS.md & USER.md
-            tools_info = acct_res["files"].get("TOOLS.md", {})
-            user_info = acct_res["files"].get("USER.md", {})
-            if tools_info.get("size", 0) > 400 and user_info.get("size", 0) > 400:
-                acct_res["drive_status"]["context"] = "FULL_CONTEXT"
-                acct_res["drive_score"] += 25
-            else:
-                acct_res["drive_status"]["context"] = "PARTIAL_OR_EMPTY"
-                acct_res["issues"].append("TOOLS.md or USER.md missing operational conventions")
-
-        except Exception as e:
-            acct_res["error"] = str(e)
-            acct_res["issues"].append(f"Connection failed: {e}")
-
-        results[acct] = acct_res
-
-    return results
-
-
-def diff_md(account: str, filename: str) -> dict:
-    """Compare an agent's container file against the shared operator template."""
-    validate_filename(filename, template_only=True)
-    local_path = SHARED_OPERATORS / filename
-    if not local_path.exists():
-        raise FileNotFoundError(f"Local template {local_path} not found")
-        
-    local_content = local_path.read_text(encoding="utf-8")
-    remote_data = read_md(account, filename)
-    remote_content = remote_data.get("text", "")
-    
-    diff = list(difflib.unified_diff(
-        remote_content.splitlines(keepends=True),
-        local_content.splitlines(keepends=True),
-        fromfile=f"{account}:{filename} (remote)",
-        tofile=f"shared/operators/{filename} (local)",
-    ))
-    
-    return {
-        "ok": True,
-        "account": account,
-        "filename": filename,
-        "identical": len(diff) == 0,
-        "diff": "".join(diff),
-        "remote_size": len(remote_content.encode("utf-8")),
-        "local_size": len(local_content.encode("utf-8")),
-    }
-
-
-def amend_md(filename: str, content: str, author: str = "operator", reason: str = "") -> dict:
-    """Amend a centralized shared operator template in shared/operators/ with safety validation and git commit."""
-    import subprocess
-    validate_filename(filename, template_only=True)
-    
-    local_path = SHARED_OPERATORS / filename
-    if not local_path.exists():
-        raise FileNotFoundError(f"Shared operator file {filename} does not exist in {SHARED_OPERATORS}")
-        
-    # Drive safety validation
-    if filename == "HEARTBEAT.md":
-        # Ensure checklist is not gutted
-        non_comment_lines = [l for l in content.splitlines() if l.strip() and not l.strip().startswith("#")]
-        checklist_items = [l for l in non_comment_lines if l.strip().startswith("- [")]
-        if not checklist_items:
-            raise ValueError("Safety rejection: amendment removes all active checklist items from HEARTBEAT.md")
-            
-    elif filename == "PROACTIVE_PREFERENCES.md":
-        if len(content.strip()) < 400:
-            raise ValueError("Safety rejection: amendment would reduce PROACTIVE_PREFERENCES.md to unconfigured state")
-
-    elif filename == "SOUL.md":
-        if "Be a guest in someone's life" in content and "AUTONOMOUS OPERATIONAL DRIVE" not in content:
-            raise ValueError("Safety rejection: amendment reverts SOUL.md to passive stock template")
-
-    old_content = local_path.read_text(encoding="utf-8")
-    local_path.write_text(content, encoding="utf-8")
-    
-    # Git auto-commit if in git repo
-    git_committed = False
-    git_hash = None
-    try:
-        commit_msg = f"amend(operators): update {filename} via {author}"
-        if reason:
-            commit_msg += f" - {reason}"
-        subprocess.run(["git", "add", str(local_path)], cwd=str(NETVM_ROOT), check=True, capture_output=True)
-        cr = subprocess.run(["git", "commit", "-m", commit_msg], cwd=str(NETVM_ROOT), capture_output=True, text=True)
-        if cr.returncode == 0:
-            git_committed = True
-            hr = subprocess.run(["git", "rev-parse", "--short", "HEAD"], cwd=str(NETVM_ROOT), capture_output=True, text=True)
-            git_hash = hr.stdout.strip()
-    except Exception:
-        pass
-
-    return {
-        "ok": True,
-        "action": "amend",
-        "filename": filename,
-        "author": author,
-        "reason": reason,
-        "bytes_written": len(content.encode("utf-8")),
-        "git_committed": git_committed,
-        "commit": git_hash,
-    }
-
-
-def append_md(filename: str, text: str, author: str = "operator", section: str = None) -> dict:
-    """Safely append an amendment or lesson to a centralized shared template."""
-    validate_filename(filename, template_only=True)
-    local_path = SHARED_OPERATORS / filename
-    if not local_path.exists():
-        raise FileNotFoundError(f"Shared operator file {filename} does not exist in {SHARED_OPERATORS}")
-        
-    current = local_path.read_text(encoding="utf-8")
-    header = f"\n\n<!-- Amendment by {author} on {time.strftime('%Y-%m-%d %H:%M:%S UTC', time.gmtime())} -->\n"
-    if section:
-        header += f"### {section}\n"
-    new_content = current.rstrip() + header + text.strip() + "\n"
-    
-    return amend_md(filename, new_content, author=author, reason=f"append {section or 'note'}")
-
-
-def pull_md(account: str, filename: str) -> dict:
-    """Pull the canonical centralized template from shared/operators/ into an agent's container."""
-    validate_filename(filename, template_only=True)
-    local_path = SHARED_OPERATORS / filename
-    if not local_path.exists():
-        raise FileNotFoundError(f"Shared operator file {filename} does not exist in {SHARED_OPERATORS}")
-    content = local_path.read_text(encoding="utf-8")
-    res = write_md(account, filename, content, overwrite=True)
-    return {
-        "ok": True,
-        "account": account,
-        "filename": filename,
-        "bytes_written": res.get("bytes_written"),
-        "message": f"Successfully pulled canonical {filename} into {account} container",
-    }
-
-
-def inject_drive(account: str, force: bool = False) -> dict:
-    """Inject high-drive operational instructions into the agent's container."""
-    updates = []
-    
-    # 1. SOUL.md
-    soul_text = (SHARED_OPERATORS / "SOUL.md").read_text(encoding="utf-8")
-    r_soul = write_md(account, "SOUL.md", soul_text, overwrite=True)
-    updates.append({"file": "SOUL.md", "bytes": r_soul["bytes_written"]})
-    
-    # 2. PROACTIVE_PREFERENCES.md
-    pro_text = (SHARED_OPERATORS / "PROACTIVE_PREFERENCES.md").read_text(encoding="utf-8")
-    r_pro = write_md(account, "PROACTIVE_PREFERENCES.md", pro_text, overwrite=True)
-    updates.append({"file": "PROACTIVE_PREFERENCES.md", "bytes": r_pro["bytes_written"]})
-    
-    # 3. HEARTBEAT.md
-    hb_text = (SHARED_OPERATORS / "HEARTBEAT.md").read_text(encoding="utf-8")
-    r_hb = write_md(account, "HEARTBEAT.md", hb_text, overwrite=True)
-    updates.append({"file": "HEARTBEAT.md", "bytes": r_hb["bytes_written"]})
-    
-    # 4. USER.md
-    user_text = (SHARED_OPERATORS / "USER.md").read_text(encoding="utf-8")
-    r_user = write_md(account, "USER.md", user_text, overwrite=True)
-    updates.append({"file": "USER.md", "bytes": r_user["bytes_written"]})
-    
-    # 5. TOOLS.md
-    tools_text = (SHARED_OPERATORS / "TOOLS.md").read_text(encoding="utf-8")
-    r_tools = write_md(account, "TOOLS.md", tools_text, overwrite=True)
-    updates.append({"file": "TOOLS.md", "bytes": r_tools["bytes_written"]})
-
-    # 6. AGENTS.md (preserve existing custom lessons if present)
-    agents_template = (SHARED_OPERATORS / "AGENTS.md").read_text(encoding="utf-8")
-    try:
-        remote_agents = read_md(account, "AGENTS.md").get("text", "")
-        if "## Lessons" in remote_agents and len(remote_agents) > len(agents_template):
-            # Extract custom lessons from remote and merge
-            custom_lessons = remote_agents.split("## Lessons", 1)[1]
-            merged_agents = agents_template.rstrip() + "\n\n## Lessons" + custom_lessons
-            r_agents = write_md(account, "AGENTS.md", merged_agents, overwrite=True)
-        else:
-            r_agents = write_md(account, "AGENTS.md", agents_template, overwrite=True)
-    except Exception:
-        r_agents = write_md(account, "AGENTS.md", agents_template, overwrite=True)
-    updates.append({"file": "AGENTS.md", "bytes": r_agents["bytes_written"]})
-
-    return {
-        "ok": True,
-        "account": account,
-        "action": "inject_drive",
-        "updates": updates,
-        "message": f"Successfully injected high-drive operator files into {account} container",
-    }
-
-
-def get_ssh_info(account: str = None) -> dict:
-    """Return SSH connection coordinates and reverse tunnel configuration."""
-    jump_host = "34.139.37.135"
-    if account:
-        entry = TUNNEL_PORTS.get(account, {"port": 2226, "terminal": 7683, "user": "hatch"})
-        port = entry["port"]
-        user = entry["user"]
-        cmd = f"ssh -o StrictHostKeyChecking=no -p {port} {user}@localhost"
-        proxy_cmd = f"ssh -o StrictHostKeyChecking=no -J super@{jump_host} -p {port} {user}@localhost"
-        return {
-            "account": account,
-            "jump_host": jump_host,
-            "port": port,
-            "container_user": user,
-            "terminal_port": entry.get("terminal"),
-            "direct_from_vm": cmd,
-            "jump_command": proxy_cmd,
-            "cat_example": f"cat file.md | {proxy_cmd} 'cat > /home/hatch/file.md'",
-        }
-    return {
-        "jump_host": jump_host,
-        "tunnels": TUNNEL_PORTS,
-    }
-
-
-def main():
-    import argparse
-    parser = argparse.ArgumentParser(description="Manage Muse agent .md files via Hatch and SSH")
-    sub = parser.add_subparsers(dest="cmd")
-    
-    p_audit = sub.add_parser("audit", help="Audit .md files and DRIVE across all agents")
-    p_audit.add_argument("accounts", nargs="*", help="Optional account filter")
-    p_audit.add_argument("--json", action="store_true", help="Output JSON")
-    
-    p_list = sub.add_parser("list", help="List container files via Hatch")
-    p_list.add_argument("account", help="Agent account")
-    p_list.add_argument("path", nargs="?", default="", help="Subdirectory path")
-    
-    p_read = sub.add_parser("read", help="Read a markdown file via Hatch")
-    p_read.add_argument("account", help="Agent account")
-    p_read.add_argument("filename", help="Filename (e.g. SOUL.md)")
-    
-    p_write = sub.add_parser("write", help="Write a markdown file via Hatch")
-    p_write.add_argument("account", help="Agent account")
-    p_write.add_argument("filename", help="Filename (e.g. SOUL.md)")
-    p_write.add_argument("--content", help="Text content to write")
-    p_write.add_argument("--file", help="Local file to copy content from")
-    
-    p_diff = sub.add_parser("diff", help="Diff remote file against shared operator template")
-    p_diff.add_argument("account", help="Agent account")
-    p_diff.add_argument("filename", help="Filename (e.g. SOUL.md)")
-    
-    p_amend = sub.add_parser("amend", help="Amend a shared operator file with validation and git commit")
-    p_amend.add_argument("filename", help="Filename (e.g. AGENTS.md, TOOLS.md)")
-    p_amend.add_argument("--content", help="New content")
-    p_amend.add_argument("--file", help="File with new content")
-    p_amend.add_argument("--author", default="operator", help="Author of amendment")
-    p_amend.add_argument("--reason", default="", help="Reason for amendment")
-
-    p_append = sub.add_parser("append", help="Safely append a note or lesson to a shared operator file")
-    p_append.add_argument("filename", help="Filename (e.g. AGENTS.md)")
-    p_append.add_argument("text", help="Text to append")
-    p_append.add_argument("--author", default="operator", help="Author of amendment")
-    p_append.add_argument("--section", default=None, help="Optional section header")
-
-    p_pull = sub.add_parser("pull", help="Pull canonical shared file into an agent's container")
-    p_pull.add_argument("account", help="Agent account")
-    p_pull.add_argument("filename", help="Filename (e.g. SOUL.md)")
-
-    p_drive = sub.add_parser("inject-drive", help="Inject high-drive operator files into agent")
-    p_drive.add_argument("account", help="Agent account")
-    p_drive.add_argument("--force", action="store_true", help="Force overwrite")
-    
-    p_sync_all = sub.add_parser("sync-all", help="Inject high-drive files across all active agents")
-    
-    p_ssh = sub.add_parser("ssh-info", help="Get SSH tunnel dial-in information")
-    p_ssh.add_argument("account", nargs="?", help="Optional agent account")
-
-    args = parser.parse_args()
-    if not args.cmd:
-        parser.print_help()
-        sys.exit(1)
-        
-    if args.cmd == "audit":
-        res = audit_agents(args.accounts or None)
-        if args.json:
-            print(json.dumps(res, indent=2))
-        else:
-            print(f"\n{'='*70}\nMUSE AGENT .MD & DRIVE AUDIT REPORT\n{'='*70}")
-            for acct, d in res.items():
-                if not d.get("connected"):
-                    print(f"\n[AGENT {acct.upper()}] ✗ Connection failed: {d.get('error')}")
-                    continue
-                score = d.get("drive_score", 0)
-                status_color = "HIGH DRIVE" if score >= 75 else ("PARTIAL" if score >= 50 else "LOW DRIVE / STALE")
-                print(f"\n[AGENT {acct.upper()}] DRIVE Score: {score}/100 ({status_color})  VM: {d.get('vm_id', 'unknown')}")
-                for fname, finfo in d.get("files", {}).items():
-                    ex = "✓" if finfo.get("exists") else "✗"
-                    sz = f"{finfo.get('size', 0):6} bytes"
-                    mod = (finfo.get("modified") or "")[:19]
-                    print(f"  {ex} {fname:24} {sz}  {mod}")
-                if d.get("issues"):
-                    print("  Issues:")
-                    for iss in d["issues"]:
-                        print(f"    • {iss}")
-            print(f"\n{'='*70}\n")
-            
-    elif args.cmd == "list":
-        entries = list_files(args.account, args.path)
-        print(json.dumps(entries, indent=2))
-        
-    elif args.cmd == "read":
-        r = read_md(args.account, args.filename)
-        print(r.get("text", ""))
-        
-    elif args.cmd == "write":
-        content = args.content
-        if args.file:
-            content = Path(args.file).read_text(encoding="utf-8")
-        if content is None:
-            print("Error: provide --content or --file", file=sys.stderr)
-            sys.exit(2)
-        res = write_md(args.account, args.filename, content)
-        print(json.dumps(res, indent=2))
-        
-    elif args.cmd == "diff":
-        res = diff_md(args.account, args.filename)
-        if res["identical"]:
-            print(f"{args.account}:{args.filename} matches local shared/operators/{args.filename} exactly.")
-        else:
-            print(res["diff"])
-
-    elif args.cmd == "amend":
-        content = args.content
-        if args.file:
-            content = Path(args.file).read_text(encoding="utf-8")
-        if content is None:
-            print("Error: provide --content or --file", file=sys.stderr)
-            sys.exit(2)
-        try:
-            res = amend_md(args.filename, content, author=args.author, reason=args.reason)
-            print(json.dumps(res, indent=2))
-        except Exception as e:
-            print(json.dumps({"ok": False, "error": str(e)}), indent=2)
-            sys.exit(1)
-
-    elif args.cmd == "append":
-        try:
-            res = append_md(args.filename, args.text, author=args.author, section=args.section)
-            print(json.dumps(res, indent=2))
-        except Exception as e:
-            print(json.dumps({"ok": False, "error": str(e)}), indent=2)
-            sys.exit(1)
-
-    elif args.cmd == "pull":
-        try:
-            res = pull_md(args.account, args.filename)
-            print(json.dumps(res, indent=2))
-        except Exception as e:
-            print(json.dumps({"ok": False, "error": str(e)}), indent=2)
-            sys.exit(1)
-            
-    elif args.cmd == "inject-drive":
-        res = inject_drive(args.account, force=args.force)
-        print(json.dumps(res, indent=2))
-        
-    elif args.cmd == "sync-all":
-        results = {}
-        for acct in VALID_ACCOUNTS:
-            try:
-                results[acct] = inject_drive(acct)
-                print(f"✓ Injected DRIVE into {acct}")
-            except Exception as e:
-                results[acct] = {"ok": False, "error": str(e)}
-                print(f"✗ Failed {acct}: {e}")
-                
-    elif args.cmd == "ssh-info":
-        res = get_ssh_info(args.account)
-        print(json.dumps(res, indent=2))
-
-
-if __name__ == "__main__":
-    main()

bin/box-chat-cdp.py

@@ -1,333 +0,0 @@
-#!/usr/bin/env python3
-"""box-chat-cdp.py — READ-ONLY CDP scraper for box-chat.py.
-
-Runs INSIDE the agent's netns (via netvm-exec.sh), where the agent's
-headless Chromium CDP port is reachable on 127.0.0.1. Performs only
-Runtime.evaluate reads plus benign navigation clicks (panel open, thread
-switch, restore-to-main). Never sends messages, never touches the
-composer, never clicks send.
-
-Usage:
-    box-chat-cdp.py <agent> threads
-    box-chat-cdp.py <agent> messages <thread-id>
-
-Prints exactly one JSON object to stdout. Exit 0 on success, 1 on
-failure (stdout still carries {"ok": false, ...}).
-"""
-
-import importlib.util
-import json
-import sys
-import time
-import urllib.request
-
-
-def load_accounts():
-    path = "/home/super/Projects/NetVM/bin/netvm-registry.py"
-    spec = importlib.util.spec_from_file_location("netvm_registry", path)
-    mod = importlib.util.module_from_spec(spec)
-    spec.loader.exec_module(mod)
-    accounts = {}
-    for node, rec in mod.load().items():
-        accounts[node] = (node, "http://127.0.0.1:%d/json/list" % rec["cdp_port"])
-    return accounts
-
-
-def connect(agent):
-    accounts = load_accounts()
-    if agent not in accounts:
-        raise RuntimeError("unknown agent in registry: %s" % agent)
-    _node, cdp_url = accounts[agent]
-    with urllib.request.urlopen(cdp_url, timeout=10) as r:
-        targets = json.load(r)
-    pages = [t for t in targets if t.get("type") == "page"]
-    if not pages:
-        raise RuntimeError("no page target on CDP")
-    import websocket
-    ws = websocket.create_connection(
-        pages[0]["webSocketDebuggerUrl"], timeout=60)
-    return ws
-
-
-def ev(ws, expr, await_p=False):
-    ws.send(json.dumps({
-        "id": 1, "method": "Runtime.evaluate",
-        "params": {"expression": expr, "returnByValue": True,
-                   "awaitPromise": await_p},
-    }))
-    # Drain CDP events until we get our command response (id 1).
-    # The browser can emit events (Runtime.executionContextCreated, etc.)
-    # at any time; taking the first recv() blindly returns None on a
-    # busy page (observed as transient NO_SWITCHER / "unexpected
-    # messages payload" on pip/opm 2026-10-04).
-    resp = None
-    drained = 0
-    for _ in range(50):
-        raw = ws.recv()
-        resp = json.loads(raw)
-        if resp.get("id") == 1:
-            break
-        drained += 1
-    else:
-        raise RuntimeError("CDP: no response to Runtime.evaluate (drained %d)" % drained)
-    pass  # drained count available in `drained` if needed
-    res = resp.get("result", {})
-    if res.get("subtype") == "error":
-        raise RuntimeError("JS error: %s" % str(res.get("description"))[:200])
-    return res.get("result", {}).get("value")
-
-
-TITLE_OF = """const titleOf = r => {
-  const s = r.querySelector('span[title]');
-  return s ? s.getAttribute('title').trim()
-           : ((r.innerText||'').split('\\n')[0]||'').trim();
-};"""
-
-ENSURE = """(async () => {
-  const sleep = ms => new Promise(r => setTimeout(r, ms));
-  %s
-  if (window.location.pathname === '/thread/new') {
-    window.location.href = '/'; await sleep(4000);
-  }
-  const nav = document.querySelector('[data-testid="hatch-nav-chat"]');
-  if (!nav) return 'NO_CHAT_NAV';
-  if (nav.getAttribute('aria-current') !== 'page') { nav.click(); await sleep(3000); }
-  const panelOpen = () => !!document.querySelector('[data-testid="hatch-chat-compose"]');
-  if (!panelOpen()) {
-    // Retry: the switcher may not be rendered yet if the SPA is still
-    // settling (observed transient NO_SWITCHER on pip/opm 2026-10-04).
-    let sw = null;
-    for (let k = 0; k < 4 && !sw; k++) {
-      sw = document.querySelector('[data-testid="hatch-chat-switcher-trigger"]');
-      if (!sw) await sleep(2000);
-    }
-    if (!sw) return 'NO_SWITCHER';
-    sw.click(); await sleep(2500);
-    if (!panelOpen()) return 'PANEL_CLOSED';
-  }
-  return 'OK';
-})()""" % TITLE_OF
-
-
-def op_threads(ws):
-    st = ev(ws, ENSURE, await_p=True)
-    if st != "OK":
-        raise RuntimeError("could not reach chat panel: %s" % st)
-    js = """(async () => {
-      const sleep = ms => new Promise(r => setTimeout(r, ms));
-      %s
-      const snap = [...document.querySelectorAll('[data-testid="hatch-thread-row"]')]
-        .map(r => {
-          const spans = [...r.querySelectorAll('span')];
-          return {title: titleOf(r),
-                  rel: spans.length ? (spans[spans.length-1].innerText||'').trim() : ''};
-        });
-      const ensurePanel = async () => {
-        if (document.querySelector('[data-testid="hatch-chat-compose"]')) return true;
-        for (let k = 0; k < 3; k++) {
-          const sw = document.querySelector('[data-testid="hatch-chat-switcher-trigger"]');
-          if (!sw) return false;
-          sw.click(); await sleep(3000);
-          if (document.querySelector('[data-testid="hatch-chat-compose"]')) return true;
-        }
-        return false;
-      };
-      const out = [];
-      for (const s of snap) {
-        const isMain = /^main chat$/i.test(s.title);
-        if (isMain) { out.push({id: 'main', kind: 'main', title: s.title, rel: s.rel}); continue; }
-        const panelOk = await ensurePanel();
-        if (!panelOk) { out.push({id: null, kind: 'sidechat', title: s.title, rel: s.rel, error: 'PANEL_CLOSED'}); continue; }
-        const row = [...document.querySelectorAll('[data-testid="hatch-thread-row"]')]
-          .find(r => titleOf(r) === s.title);
-        if (!row) { out.push({id: null, kind: 'sidechat', title: s.title, rel: s.rel, error: 'ROW_GONE'}); continue; }
-        row.click();
-        let id = null;
-        for (let t = 0; t < 20; t++) {
-          await sleep(500);
-          const m = window.location.href.match(/\\/thread\\/([0-9a-fA-F-]{36})/);
-          if (m) { id = m[1]; break; }
-        }
-        if (!id) {
-          const row2 = [...document.querySelectorAll('[data-testid="hatch-thread-row"]')]
-            .find(r => titleOf(r) === s.title);
-          if (row2) {
-            row2.click();
-            for (let t = 0; t < 12; t++) {
-              await sleep(500);
-              const m = window.location.href.match(/\\/thread\\/([0-9a-fA-F-]{36})/);
-              if (m) { id = m[1]; break; }
-            }
-          }
-        }
-        out.push({id, kind: 'sidechat', title: s.title, rel: s.rel});
-      }
-      return {threads: out};
-    })()""" % TITLE_OF
-    return ev(ws, js, await_p=True)
-
-
-def op_messages(ws, thread_id):
-    st = ev(ws, ENSURE, await_p=True)
-    if st != "OK":
-        raise RuntimeError("could not reach chat panel: %s" % st)
-    js = """(async () => {
-      const sleep = ms => new Promise(r => setTimeout(r, ms));
-      %s
-      const THREAD = %s;
-      const ensurePanel = async () => {
-        if (document.querySelector('[data-testid="hatch-chat-compose"]')) return true;
-        for (let k = 0; k < 3; k++) {
-          const sw = document.querySelector('[data-testid="hatch-chat-switcher-trigger"]');
-          if (!sw) return false;
-          sw.click(); await sleep(3000);
-          if (document.querySelector('[data-testid="hatch-chat-compose"]')) return true;
-        }
-        return false;
-      };
-      const uuidOf = () => {
-        const m = window.location.href.match(/\\/thread\\/([0-9a-fA-F-]{36})/);
-        return m ? m[1].toLowerCase() : null;
-      };
-      let landed = false;
-      if (THREAD === 'main') {
-        await ensurePanel();
-        const row = [...document.querySelectorAll('[data-testid="hatch-thread-row"]')]
-          .find(r => /^main chat$/i.test(titleOf(r)));
-        if (row) { row.click(); await sleep(2500); }
-        landed = /muse\\.ai\\/?$/.test(window.location.href) && !uuidOf();
-      } else {
-        // already there?
-        if (uuidOf() === THREAD.toLowerCase()) landed = true;
-        // click rows until the URL carries our thread uuid (SPA navigation,
-        // keeps the JS context alive unlike location.href assignment)
-        for (let i = 0; i < 12 && !landed; i++) {
-          await ensurePanel();
-          const rows = [...document.querySelectorAll('[data-testid="hatch-thread-row"]')];
-          if (!rows.length) break;
-          const row = rows[i %% rows.length];
-          row.click();
-          for (let t = 0; t < 14; t++) {
-            await sleep(500);
-            if (uuidOf() === THREAD.toLowerCase()) { landed = true; break; }
-            if (uuidOf()) break; // navigated somewhere else; try next row
-          }
-        }
-        if (!landed) {
-          // fallback: SPA history navigation (no full page load)
-          window.history.pushState({}, '', '/thread/' + THREAD);
-          window.dispatchEvent(new PopStateEvent('popstate'));
-          await sleep(5000);
-          landed = uuidOf() === THREAD.toLowerCase();
-        }
-      }
-      if (!landed) return {error: 'THREAD_NOT_FOUND'};
-      await sleep(2500);
-      const sc = document.getElementById('hatch-chat-scroll');
-      if (sc) {
-        for (let i = 0; i < 3; i++) { sc.scrollTop = 0; await sleep(1500); }
-        sc.scrollTop = sc.scrollHeight; await sleep(800);
-      }
-      const els = [...document.querySelectorAll('[data-message-id]')];
-      const messages = els.map(m => {
-        const id = m.getAttribute('data-message-id');
-        const ps = [...m.querySelectorAll('p')].map(p => (p.innerText||'').trim()).filter(Boolean);
-        let text = ps.join('\\n');
-        if (!text) text = (m.innerText||'').replace(/^(Assistant message:|User message:)\\s*/, '').trim();
-        const t = m.querySelector('time');
-        return {id,
-                author: id.indexOf('assistant-msg') === 0 ? 'assistant' : 'user',
-                text,
-                ts: t ? (t.getAttribute('datetime') || t.innerText || null) : null};
-      });
-      return {url: window.location.href, messages};
-    })()""" % (TITLE_OF, json.dumps(thread_id))
-    data = ev(ws, js, await_p=True)
-    if not isinstance(data, dict) or "messages" not in data:
-        if isinstance(data, dict) and data.get("error") == "THREAD_NOT_FOUND":
-            raise RuntimeError("THREAD_NOT_FOUND: no such thread for this agent")
-        raise RuntimeError("unexpected messages payload")
-    return data
-
-
-def restore_main(ws):
-    try:
-        ev(ws, """(() => {
-          %s
-          const row = [...document.querySelectorAll('[data-testid="hatch-thread-row"]')]
-            .find(r => /^main chat$/i.test(titleOf(r)));
-          if (row) row.click();
-          return 'ok';
-        })()""" % TITLE_OF)
-    except Exception:
-        pass
-
-
-def main(argv):
-    if len(argv) < 3:
-        print(json.dumps({"ok": False, "code": "BAD_ARGS",
-                          "error": "usage: box-chat-cdp.py <agent> threads|messages [thread-id]"}))
-        return 1
-    agent, op = argv[1], argv[2]
-    ws = None
-    try:
-        # CDP evaluate can resolve null when the page is mid-navigation
-        # (agent actively using browser). Reconnect fresh on each retry
-        # so we attach to the current page, not a stale JS context.
-        # Observed 2026-10-04: opm's browser navigates during reads.
-        data, last_err = None, None
-        for attempt in range(3):
-            try:
-                if ws is not None:
-                    try:
-                        ws.close()
-                    except Exception:
-                        pass
-                ws = connect(agent)
-                if op == "threads":
-                    data = op_threads(ws)
-                    ok = isinstance(data, dict) and isinstance(
-                        data.get("threads"), list)
-                elif op == "messages":
-                    if len(argv) < 4:
-                        raise RuntimeError("messages requires thread-id")
-                    data = op_messages(ws, argv[3])
-                    ok = isinstance(data, dict) and isinstance(
-                        data.get("messages"), list)
-                else:
-                    raise RuntimeError("unknown op: %s" % op)
-                if ok:
-                    break
-                last_err = "empty CDP result"
-                data = None
-            except RuntimeError as e:
-                last_err = str(e)
-                data = None
-            time.sleep(3)
-        if data is None:
-            raise RuntimeError(last_err or "CDP returned no usable data")
-        if op == "threads":
-            print(json.dumps({"ok": True, "agent": agent,
-                              "threads": data["threads"]}))
-        else:
-            print(json.dumps({"ok": True, "agent": agent, "url": data["url"],
-                              "messages": data["messages"]}))
-        return 0
-    except Exception as e:
-        print(json.dumps({"ok": False, "code": "CDP_ERROR",
-                          "error": str(e)[:300]}))
-        return 1
-    finally:
-        if ws is not None:
-            try:
-                restore_main(ws)
-            except Exception:
-                pass
-            try:
-                ws.close()
-            except Exception:
-                pass
-
-
-if __name__ == "__main__":
-    sys.exit(main(sys.argv))

bin/box-chat.py

@@ -1,380 +0,0 @@
-#!/usr/bin/env python3
-"""box-chat.py — allowlisted bl helper for Box thread oversight (READ-ONLY).
-
-The board server (VM) never scrapes browsers directly. All thread reads go
-through this helper, invoked as:
-
-    /home/super/Projects/NetVM/bin/box-chat.py thread-list <agent> [--kind K] [--limit N]
-    /home/super/Projects/NetVM/bin/box-chat.py thread-messages <agent> <thread-id> [--limit N] [--before MSGID]
-
-Security properties (mirrors box-ctl.py):
-- Fixed verb set; every argument validated before acting.
-- <agent> must be a known node (muse, pip, 646, opm); anything else exits
-  before any netns/SSH/CDP work.
-- <thread-id> must match ^[a-zA-Z0-9-]{1,64}$ or be the literal "main".
-- READ-ONLY by construction: the CDP driver (box-chat-cdp.py) only runs
-  Runtime.evaluate reads plus benign navigation clicks. No sends, no
-  composer interaction, no shell=True anywhere. All subprocess calls use
-  argv lists.
-- Every invocation audit-logged to box-chat.jsonl with caller identity.
-
-Output: JSON to stdout ({"ok": true, ...} or {"ok": false, ...}),
-exit 0 on success, nonzero on failure.
-"""
-
-import json
-import os
-import re
-import subprocess
-import sys
-from datetime import datetime, timedelta, timezone
-from pathlib import Path
-
-NETVM_ROOT = Path("/home/super/Projects/NetVM")
-BIN = NETVM_ROOT / "bin"
-CDP_DRIVER = BIN / "box-chat-cdp.py"
-NETVM_EXEC = BIN / "netvm-exec.sh"
-CHAT_LOG = NETVM_ROOT / "box-chat.jsonl"
-DM_LOG = NETVM_ROOT / "dm-log.jsonl"
-
-VALID_AGENTS = {"muse", "pip", "646", "opm"}
-VALID_KINDS = {"main", "sidechat", "dm", "all"}
-AGENT_RE = re.compile(r"^[a-z0-9-]{1,64}$")
-THREAD_RE = re.compile(r"^[a-zA-Z0-9-]{1,64}$")
-MSGID_RE = re.compile(r"^[a-zA-Z0-9-]{1,128}$")
-
-REL_MONTHS = {m: i + 1 for i, m in enumerate(
-    ["jan", "feb", "mar", "apr", "may", "jun",
-     "jul", "aug", "sep", "oct", "nov", "dec"])}
-
-
-def utcnow():
-    return datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
-
-
-def out(ok, **kw):
-    payload = {"ok": ok}
-    payload.update(kw)
-    print(json.dumps(payload))
-
-
-def fail(code, error, detail=None, exit_code=1):
-    payload = {"ok": False, "code": code, "error": error}
-    if detail is not None:
-        payload["detail"] = detail
-    print(json.dumps(payload))
-    sys.exit(exit_code)
-
-
-def audit(action, agent=None, name=None, kind=None):
-    """Append invocation record to the bl-side log."""
-    try:
-        entry = {
-            "ts": utcnow(),
-            "action": action,
-            "agent": agent,
-            "name": name,
-            "kind": kind,
-            "caller": os.environ.get("BOX_CALLER", "unknown"),
-        }
-        with open(CHAT_LOG, "a") as f:
-            f.write(json.dumps(entry) + "\n")
-    except Exception:
-        pass  # audit failure must not break the action
-
-
-def check_agent(agent):
-    if not agent or agent not in VALID_AGENTS:
-        fail("BAD_AGENT", "agent must be one of %s" % sorted(VALID_AGENTS),
-             {"field": "agent", "value": agent})
-    return agent
-
-
-def check_thread_id(tid):
-    if tid != "main" and not (tid and THREAD_RE.match(tid)):
-        fail("BAD_THREAD", "thread-id must be 'main' or match ^[a-zA-Z0-9-]{1,64}$",
-             {"field": "thread-id", "value": tid})
-    return tid
-
-
-def check_limit(val, default):
-    if val is None:
-        return default
-    try:
-        n = int(val)
-    except (TypeError, ValueError):
-        fail("BAD_LIMIT", "limit must be an integer 1-200", {"value": val})
-    if not 1 <= n <= 200:
-        fail("BAD_LIMIT", "limit must be an integer 1-200", {"value": val})
-    return n
-
-
-def run_cdp(agent, op, args, timeout):
-    """Run the CDP driver inside the agent's netns. argv only, no shell."""
-    cmd = [str(NETVM_EXEC), agent, "--", sys.executable,
-           str(CDP_DRIVER), agent, op] + args
-    try:
-        r = subprocess.run(cmd, capture_output=True, text=True,
-                           timeout=timeout)
-    except subprocess.TimeoutExpired:
-        fail("CDP_TIMEOUT", "CDP driver timed out in netns",
-             {"agent": agent, "op": op})
-    if r.returncode != 0 and not r.stdout.strip():
-        fail("CDP_ERROR", "CDP driver failed",
-             {"stderr": (r.stderr or "")[-300:]})
-    try:
-        data = json.loads(r.stdout.strip())
-    except Exception:
-        fail("CDP_ERROR", "CDP driver returned non-JSON",
-             {"stdout": r.stdout[-300:], "stderr": (r.stderr or "")[-300:]})
-    if not data.get("ok"):
-        code = data.get("code", "CDP_ERROR")
-        if "THREAD_NOT_FOUND" in str(data.get("error", "")):
-            code = "THREAD_NOT_FOUND"
-        fail(code, data.get("error", "cdp failed"))
-    return data
-
-
-def parse_rel_time(rel):
-    """Panel relative times ('just now', '4m', '2h', '3d', 'Oct 3') ->
-    (approx ISO8601, True). Returns (None, False) when unparseable."""
-    now = datetime.now(timezone.utc)
-    rel = (rel or "").strip().lower()
-    if rel in ("just now", "now"):
-        return now.strftime("%Y-%m-%dT%H:%M:%SZ"), True
-    m = re.match(r"^(\d+)\s*m(in(ute)?s?)?$", rel)
-    if m:
-        return (now - timedelta(minutes=int(m.group(1)))).strftime("%Y-%m-%dT%H:%M:%SZ"), True
-    m = re.match(r"^(\d+)\s*h((ou)?rs?)?$", rel)
-    if m:
-        return (now - timedelta(hours=int(m.group(1)))).strftime("%Y-%m-%dT%H:%M:%SZ"), True
-    m = re.match(r"^(\d+)\s*d(ays?)?$", rel)
-    if m:
-        return (now - timedelta(days=int(m.group(1)))).strftime("%Y-%m-%dT%H:%M:%SZ"), True
-    m = re.match(r"^([a-z]{3})\s+(\d{1,2})$", rel)
-    if m and m.group(1) in REL_MONTHS:
-        dt = now.replace(month=REL_MONTHS[m.group(1)], day=int(m.group(2)),
-                         hour=12, minute=0, second=0, microsecond=0)
-        if dt > now:
-            dt = dt.replace(year=dt.year - 1)
-        return dt.strftime("%Y-%m-%dT%H:%M:%SZ"), True
-    return None, False
-
-
-def dm_conversations(agent):
-    """DM conversations involving <agent>, synthesized from dm-log.jsonl.
-
-    Real timestamps and counts; bodies are not stored in the log (by design)
-    — message text for these threads is the wire tag, and full bodies live
-    in the target chat's messages.
-    """
-    convos = {}
-    try:
-        with open(DM_LOG) as f:
-            for line in f:
-                line = line.strip()
-                if not line:
-                    continue
-                try:
-                    e = json.loads(line)
-                except Exception:
-                    continue
-                if e.get("type") not in ("sent", "send_done"):
-                    continue
-                frm, to = e.get("agent"), e.get("to")
-                if not frm or not to:
-                    continue
-                if agent not in (frm, to):
-                    continue
-                key = tuple(sorted([frm, to]))
-                c = convos.setdefault(key, {"count": 0, "last_ts": "",
-                                            "entries": []})
-                c["count"] += 1
-                if e.get("ts", "") > c["last_ts"]:
-                    c["last_ts"] = e["ts"]
-                c["entries"].append(e)
-    except FileNotFoundError:
-        return []
-    out = []
-    for (a, b), c in sorted(convos.items()):
-        out.append({
-            "id": "dm-%s-%s" % (a, b),
-            "kind": "dm",
-            "title": "dm:%s:%s" % (a, b),
-            "participants": [a, b],
-            "last_message_at": c["last_ts"] or None,
-            "last_message_approx": False,
-            "message_count": c["count"],
-        })
-    return out
-
-
-def dm_thread_messages(agent, thread_id):
-    """Messages for a dm-<a>-<b> thread, from dm-log.jsonl (complete log)."""
-    parts = thread_id[3:].split("-")
-    if len(parts) != 2:
-        fail("THREAD_NOT_FOUND", "no such DM thread: %s" % thread_id)
-    a, b = parts
-    if agent not in (a, b):
-        fail("THREAD_NOT_FOUND", "no such DM thread: %s" % thread_id)
-    msgs = []
-    try:
-        with open(DM_LOG) as f:
-            for line in f:
-                line = line.strip()
-                if not line:
-                    continue
-                try:
-                    e = json.loads(line)
-                except Exception:
-                    continue
-                if e.get("type") not in ("sent", "send_done"):
-                    continue
-                frm, to = e.get("agent"), e.get("to")
-                if not frm or not to:
-                    continue
-                if tuple(sorted([frm, to])) != (a, b):
-                    continue
-                sender = e.get("agent")
-                msgs.append({
-                    "id": str(e.get("id", "")),
-                    "from": {"role": "agent", "name": sender},
-                    "text": "[from:%s] [id:%s] -> %s" % (
-                        sender, e.get("id"), e.get("target", "?")),
-                    "ts": e.get("ts"),
-                    "target": e.get("target"),
-                    "verified": e.get("verified"),
-                })
-    except FileNotFoundError:
-        pass
-    msgs.sort(key=lambda m: m.get("ts") or "")
-    # de-dupe send_done/sent pairs on DM id, keep the richer record
-    seen = {}
-    for m in msgs:
-        prev = seen.get(m["id"])
-        if prev is None or (m.get("verified") and not prev.get("verified")):
-            seen[m["id"]] = m
-    msgs = sorted(seen.values(), key=lambda m: m.get("ts") or "")
-    return msgs
-
-
-def act_thread_list(agent, kind, limit):
-    threads = []
-    if kind in ("main", "sidechat", "all"):
-        data = run_cdp(agent, "threads", [], timeout=240)
-        for t in data.get("threads", [])[:limit]:
-            if kind != "all" and t.get("kind") != kind:
-                continue
-            ts, approx = parse_rel_time(t.get("rel", ""))
-            threads.append({
-                "id": t.get("id"),
-                "kind": t.get("kind"),
-                "title": t.get("title"),
-                "participants": [agent, "human"],
-                "last_message_at": ts,
-                "last_message_approx": approx,
-                "message_count": None,  # list is a panel scrape; counts need a thread open
-            })
-    if kind in ("dm", "all"):
-        threads.extend(dm_conversations(agent))
-    audit("thread-list", agent=agent, kind=kind)
-    out(True, agent=agent, kind=kind, threads=threads,
-        fetched_at=utcnow())
-
-
-def act_thread_messages(agent, thread_id, limit, before):
-    if thread_id.startswith("dm-"):
-        msgs = dm_thread_messages(agent, thread_id)
-        thread = {"id": thread_id, "kind": "dm",
-                  "title": "dm:%s" % thread_id[3:].replace("-", ":"),
-                  "participants": thread_id[3:].split("-")}
-    else:
-        data = run_cdp(agent, "messages", [thread_id], timeout=150)
-        raw = data.get("messages", [])
-        thread = {"id": thread_id,
-                  "kind": "main" if thread_id == "main" else "sidechat",
-                  "participants": [agent, "human"]}
-        msgs = [{
-            "id": m.get("id"),
-            "from": ({"role": "agent", "name": agent}
-                     if m.get("author") == "assistant"
-                     else {"role": "human", "name": "human"}),
-            "text": m.get("text", ""),
-            "ts": m.get("ts"),
-        } for m in raw]
-    if before:
-        if not MSGID_RE.match(before):
-            fail("BAD_CURSOR", "before must match ^[a-zA-Z0-9-]{1,128}$",
-                 {"value": before})
-        idx = next((i for i, m in enumerate(msgs) if m["id"] == before), None)
-        if idx is None:
-            fail("BAD_CURSOR", "no message with that id in loaded window",
-                 {"value": before})
-        msgs = msgs[:idx]
-    older = len(msgs)
-    if len(msgs) > limit:
-        msgs = msgs[-limit:]
-    next_before = msgs[0]["id"] if older > len(msgs) and msgs else None
-    audit("thread-messages", agent=agent, name=thread_id)
-    out(True, agent=agent, thread=thread, messages=msgs,
-        next_before=next_before, loaded_count=older, fetched_at=utcnow())
-
-
-USAGE = """usage: box-chat.py <action> [args]
-
-  thread-list <agent> [--kind main|sidechat|dm|all] [--limit N]
-  thread-messages <agent> <thread-id> [--limit N] [--before MSGID]
-
-read-only. <agent> is one of: muse, pip, 646, opm."""
-
-
-def parse_flags(rest, names):
-    """Parse [--flag value] pairs; returns (positionals, {flag: value})."""
-    pos, flags = [], {}
-    i = 0
-    while i < len(rest):
-        tok = rest[i]
-        if tok.startswith("--") and tok[2:] in names:
-            if i + 1 >= len(rest):
-                fail("BAD_ARGS", "flag %s needs a value" % tok)
-            flags[tok[2:]] = rest[i + 1]
-            i += 2
-        elif tok.startswith("--"):
-            fail("BAD_ARGS", "unknown flag: %s" % tok)
-        else:
-            pos.append(tok)
-            i += 1
-    return pos, flags
-
-
-def main(argv):
-    if len(argv) < 2:
-        print(USAGE, file=sys.stderr)
-        sys.exit(2)
-    action = argv[1]
-
-    if action == "thread-list":
-        pos, flags = parse_flags(argv[2:], {"kind", "limit"})
-        if len(pos) != 1:
-            fail("BAD_ARGS", "usage: thread-list <agent> [--kind K] [--limit N]")
-        agent = check_agent(pos[0])
-        kind = flags.get("kind", "all")
-        if kind not in VALID_KINDS:
-            fail("BAD_ARGS", "kind must be one of %s" % sorted(VALID_KINDS))
-        act_thread_list(agent, kind, check_limit(flags.get("limit"), 50))
-    elif action == "thread-messages":
-        pos, flags = parse_flags(argv[2:], {"limit", "before"})
-        if len(pos) != 2:
-            fail("BAD_ARGS",
-                 "usage: thread-messages <agent> <thread-id> [--limit N] [--before MSGID]")
-        agent = check_agent(pos[0])
-        tid = check_thread_id(pos[1])
-        act_thread_messages(agent, tid, check_limit(flags.get("limit"), 50),
-                            flags.get("before"))
-    else:
-        print(USAGE, file=sys.stderr)
-        fail("BAD_ARGS", "unknown action: %s" % action)
-
-
-if __name__ == "__main__":
-    main(sys.argv)

bin/box-onboard-tui.py

@@ -1,720 +0,0 @@
-#!/usr/bin/env python3
-"""box-onboard-tui.py — Dedicated interactive TUI for NetVM Onboard Connects, Tmux Workers, and Auto-Approvals.
-
-Features 4 bridged surfaces:
-  [1 / F1] Onboard Connects:
-           Live inventory of fleet nodes & client onboarding pipelines,
-           invite codes, token feeding urgency, OTP verification, and salvage dispatch.
-  [2 / F2] Tmux Workers & Tally:
-           Multi-socket worker inventory (/tmp/tmux-1000/default, lte, muse.sock, netns socks),
-           active panes, live pane scrollback preview, worker spawning, and session killing.
-  [3 / F3] Auto-Approvals & Regex Matcher:
-           Master auto-approval toggle, per-agent policies, terminal regex rule engine
-           (Muse Code runs, A/B/C choices, 1/2 menus, y/n confirmations), and interactive regex tester.
-  [4 / F4] Box Surface & Logs:
-           Surface link with https://box.muse-dev.online/, real-time audit log stream,
-           and search/filter capabilities.
-"""
-
-from __future__ import annotations
-
-import curses
-import json
-import os
-import re
-import subprocess
-import sys
-import time
-from dataclasses import asdict
-from datetime import datetime, timezone
-from pathlib import Path
-from typing import Any, Dict, List, Optional, Tuple
-
-REPO_ROOT = Path(__file__).resolve().parent.parent
-BIN_DIR = REPO_ROOT / "bin"
-sys.path.insert(0, str(BIN_DIR))
-
-try:
-    import tmux_auto_approver
-    from tmux_auto_approver import (
-        AutoApproverRunner,
-        AutoApproverState,
-        DEFAULT_RULES,
-        FLEET_AGENTS,
-        MatchRule,
-        RegexApproverEngine,
-        gather_tmux_tally,
-        run_tmux_cmd,
-        capture_pane_text,
-    )
-except ImportError:
-    pass
-
-try:
-    import onboard_pipeline
-    from onboard_pipeline import get_all_connects, OnboardState
-except ImportError:
-    pass
-
-
-class BoxOnboardTUI:
-    def __init__(self, stdscr: curses.window):
-        self.stdscr = stdscr
-        self.current_tab = 0  # 0: Onboard, 1: Tmux, 2: Auto-Approvals, 3: Surface & Logs
-        self.tabs = [
-            "1: ONBOARD CONNECTS",
-            "2: TMUX WORKERS & TALLY",
-            "3: AUTO-APPROVALS & REGEX",
-            "4: SURFACE & LOGS",
-        ]
-
-        # Curses initialization
-        try:
-            curses.curs_set(0)
-        except Exception:
-            pass
-        self.stdscr.nodelay(True)
-        self.stdscr.keypad(True)
-
-        if hasattr(curses, "set_escdelay"):
-            try:
-                curses.set_escdelay(25)
-            except Exception:
-                pass
-
-        try:
-            curses.mousemask(curses.ALL_MOUSE_EVENTS | curses.REPORT_MOUSE_POSITION)
-            if hasattr(curses, "mouseinterval"):
-                curses.mouseinterval(0)
-            sys.stdout.write("\033[?1000h\033[?1002h\033[?1006h\033[?2004h")
-            sys.stdout.flush()
-        except Exception:
-            pass
-
-        self._init_colors()
-
-        # Shared State
-        self.approver_state = AutoApproverState.load()
-        self.tally = gather_tmux_tally(self.approver_state)
-        self.connects: List[Dict[str, Any]] = []
-        self._refresh_connects()
-
-        # Selection indices
-        self.sel_connect_idx = 0
-        self.sel_pane_idx = 0
-        self.sel_rule_idx = 0
-        self.sel_log_scroll = 0
-
-        # UI State & Modals
-        self.modal: Optional[str] = None  # "spawn_worker", "submit_otp", "test_regex", "help"
-        self.modal_input_buf = ""
-        self.modal_input_cursor = 0
-        self.toast_msg = "Welcome to Box Onboard & Tmux Console. Press ? for help."
-        self.toast_level = "info"
-        self.toast_time = time.time() + 4.0
-
-        # Interactive Regex Matcher state (Tab 3)
-        self.test_text_buf = "Would you like to run the following\n\n  $ box fleet status\n\n› 1. Yes, proceed (y)\n  2. No, and tell Muse Code what to do"
-        self.test_match_verdict: Optional[Dict[str, Any]] = None
-        self._eval_test_match()
-
-        # Runner instance for one-shot runs
-        self.runner = AutoApproverRunner()
-        self.last_auto_poll = 0.0
-
-    def _init_colors(self) -> None:
-        try:
-            curses.start_color()
-            curses.use_default_colors()
-            curses.init_pair(1, curses.COLOR_CYAN, -1)     # Accent / Info
-            curses.init_pair(2, curses.COLOR_YELLOW, -1)   # Warning / Agent
-            curses.init_pair(3, curses.COLOR_GREEN, -1)    # Success / Active
-            curses.init_pair(4, curses.COLOR_RED, -1)      # Error / Blocked
-            curses.init_pair(5, curses.COLOR_MAGENTA, -1)  # Special / Category
-            curses.init_pair(6, curses.COLOR_BLACK, curses.COLOR_CYAN)    # Header Selected
-            curses.init_pair(7, curses.COLOR_BLACK, curses.COLOR_WHITE)   # Selected Row
-            curses.init_pair(8, curses.COLOR_BLACK, curses.COLOR_YELLOW)  # Warning Banner
-            curses.init_pair(9, curses.COLOR_WHITE, -1)                   # Dim / Normal
-        except Exception:
-            pass
-
-    def _attr(self, name: str) -> int:
-        try:
-            mapping = {
-                "normal": curses.color_pair(0),
-                "cyan": curses.color_pair(1) | curses.A_BOLD,
-                "yellow": curses.color_pair(2) | curses.A_BOLD,
-                "green": curses.color_pair(3) | curses.A_BOLD,
-                "red": curses.color_pair(4) | curses.A_BOLD,
-                "magenta": curses.color_pair(5) | curses.A_BOLD,
-                "head_sel": curses.color_pair(6) | curses.A_BOLD,
-                "row_sel": curses.color_pair(7) | curses.A_BOLD,
-                "warn_banner": curses.color_pair(8) | curses.A_BOLD,
-                "dim": curses.color_pair(9) | curses.A_DIM,
-            }
-            return mapping.get(name, 0)
-        except Exception:
-            return 0
-
-    def set_toast(self, msg: str, level: str = "info") -> None:
-        self.toast_msg = msg
-        self.toast_level = level
-        self.toast_time = time.time() + 4.0
-
-    def _refresh_connects(self) -> None:
-        try:
-            self.connects = get_all_connects(fast=True)
-        except Exception:
-            self.connects = []
-
-    def _refresh_tally(self) -> None:
-        self.approver_state = AutoApproverState.load()
-        self.tally = gather_tmux_tally(self.approver_state)
-
-    def _eval_test_match(self) -> None:
-        engine = RegexApproverEngine([MatchRule(**r) for r in self.approver_state.rules])
-        v = engine.evaluate(self.test_text_buf)
-        self.test_match_verdict = {
-            "matched": v.matched,
-            "rule_name": v.rule_name,
-            "key": v.key,
-            "category": v.category,
-            "press_enter": v.press_enter,
-            "reason": v.reason,
-            "is_blocked": v.is_blocked,
-            "blocked_reason": v.blocked_reason,
-            "excerpt": v.excerpt,
-        }
-
-    # =================================================================
-    # Render Helpers
-    # =================================================================
-
-    def safe_addstr(self, y: int, x: int, text: str, attr: int = 0) -> None:
-        h, w = self.stdscr.getmaxyx()
-        if 0 <= y < h and 0 <= x < w:
-            avail = max(0, w - x - 1)
-            try:
-                self.stdscr.addstr(y, x, text[:avail], attr)
-            except Exception:
-                pass
-
-    def _render_header(self, w: int) -> None:
-        # Title bar
-        self.safe_addstr(0, 0, " " * w, self._attr("header_sel"))
-        title = " 󰢹 BOX ONBOARD & TMUX CONSOLE  [https://box.muse-dev.online/] "
-        self.safe_addstr(0, 1, title, self._attr("header_sel"))
-
-        # Master auto-approve badge in header
-        master_tag = " [● AUTO-APPROVE: ON] " if self.approver_state.global_enabled else " [○ AUTO-APPROVE: OFF] "
-        m_attr = self._attr("green") if self.approver_state.global_enabled else self._attr("warn_banner")
-        self.safe_addstr(0, max(len(title) + 2, w - len(master_tag) - 2), master_tag, m_attr)
-
-        # Tab navigation bar
-        self.safe_addstr(1, 0, " " * w, self._attr("dim"))
-        col = 1
-        for idx, tab_name in enumerate(self.tabs):
-            is_cur = (idx == self.current_tab)
-            pill = f" [{tab_name}] "
-            attr = self._attr("header_sel") if is_cur else self._attr("dim")
-            self.safe_addstr(1, col, pill, attr)
-            col += len(pill) + 2
-
-        self.safe_addstr(2, 0, "─" * w, self._attr("dim"))
-
-    def _render_footer(self, h: int, w: int) -> None:
-        y = h - 2
-        self.safe_addstr(y, 0, "─" * w, self._attr("dim"))
-
-        # Toast or Hints
-        if time.time() < self.toast_time:
-            attr = self._attr("green") if self.toast_level == "success" else (self._attr("warn_banner") if self.toast_level == "warn" else self._attr("cyan"))
-            self.safe_addstr(y + 1, 1, f" 󰋼 {self.toast_msg} ", attr)
-        else:
-            if self.current_tab == 0:
-                hints = " [ONBOARD] 1-4: Tabs  j/k: Select  n: New Onboard  o: Submit OTP  s: Salvage WO  r: Refresh  ?: Help  q: Quit"
-            elif self.current_tab == 1:
-                hints = " [TMUX] 1-4: Tabs  j/k: Select  t: Toggle Auto-Approve  n: Spawn  k: Kill  p: Prune  Enter: Full Tail  q: Quit"
-            elif self.current_tab == 2:
-                hints = " [RULES] 1-4: Tabs  Space/a: Toggle Master  m: Test Matcher  j/k: Rules  o: Trigger Once  q: Quit"
-            else:
-                hints = " [LOGS] 1-4: Tabs  j/k: Scroll  e: Sync Box API  r: Refresh  q: Quit"
-            self.safe_addstr(y + 1, 1, hints[:w - 2], self._attr("dim"))
-
-    # =================================================================
-    # Tab 1: Onboard Connects
-    # =================================================================
-
-    def _render_tab_onboard(self, h: int, w: int) -> None:
-        start_y = 3
-        max_rows = h - 7
-
-        self.safe_addstr(start_y, 2, "ACTIVE FLEET AGENTS & CLIENT ONBOARDING CONNECTS", self._attr("cyan"))
-        self.safe_addstr(start_y + 1, 2, "─" * (w - 4), self._attr("dim"))
-
-        hdr = f"  {'NODE':<8} {'TYPE':<16} {'STAGE / STATUS':<18} {'CDP':<8} {'INVITE':<10} {'ROLE / DETAIL'}"
-        self.safe_addstr(start_y + 2, 2, hdr, self._attr("bold"))
-        self.safe_addstr(start_y + 3, 2, "─" * (w - 4), self._attr("dim"))
-
-        if not self.connects:
-            self.safe_addstr(start_y + 4, 4, "(No onboard records found. Press 'n' to initiate client onboarding)", self._attr("dim"))
-            return
-
-        for idx, c in enumerate(self.connects[:max_rows]):
-            row_y = start_y + 4 + idx
-            is_sel = (idx == self.sel_connect_idx)
-            node = c.get("node", "")
-            t_str = c.get("type", "")
-            st_str = c.get("stage", c.get("status", ""))
-            cdp = str(c.get("cdp_port") or "-")
-            code = c.get("invite_code") or "-"
-            role = c.get("role") or c.get("detail") or c.get("email") or ""
-
-            line = f"  {node:<8} {t_str:<16} {st_str:<18} {cdp:<8} {code:<10} {role}"
-            attr = self._attr("selected") if is_sel else (self._attr("green") if "active" in st_str or "completed" in st_str else self._attr("normal"))
-            self.safe_addstr(row_y, 2, " " * (w - 4), attr if is_sel else 0)
-            self.safe_addstr(row_y, 2, line, attr)
-
-    # =================================================================
-    # Tab 2: Tmux Workers & Tally
-    # =================================================================
-
-    def _render_tab_tmux(self, h: int, w: int) -> None:
-        start_y = 3
-        split_h = max(6, (h - 6) // 2)
-
-        # Header summary
-        summary = f"TMUX WORKER TALLY — {self.tally.total_sessions} Sessions · {self.tally.total_panes} Panes · {self.tally.active_workers} Active across {self.tally.total_sockets} Sockets"
-        self.safe_addstr(start_y, 2, summary, self._attr("cyan"))
-
-        hdr = f"  {'Socket':<18} {'Session':<14} {'Pane':<6} {'PID':<8} {'Agent':<6} {'Cmd':<16} {'Auto-Approve'}"
-        self.safe_addstr(start_y + 1, 2, hdr, self._attr("bold"))
-        self.safe_addstr(start_y + 2, 2, "─" * (w - 4), self._attr("dim"))
-
-        panes = self.tally.panes
-        table_rows = split_h - 3
-
-        if not panes:
-            self.safe_addstr(start_y + 3, 4, "(No active tmux sessions found)", self._attr("dim"))
-        else:
-            for idx, p in enumerate(panes[:table_rows]):
-                row_y = start_y + 3 + idx
-                is_sel = (idx == self.sel_pane_idx)
-                sock_short = os.path.basename(p.socket)
-                auto_str = "[AUTO: ON]" if p.auto_approve else "[AUTO: OFF]"
-                auto_attr = self._attr("green") if p.auto_approve else self._attr("dim")
-
-                line = f"  {sock_short:<18} {p.session[:13]:<14} {p.pane_id:<6} {p.pane_pid:<8} {p.agent_node:<6} {p.current_command[:15]:<16} {auto_str}"
-                attr = self._attr("selected") if is_sel else self._attr("normal")
-                self.safe_addstr(row_y, 2, " " * (w - 4), attr if is_sel else 0)
-                self.safe_addstr(row_y, 2, line, attr)
-
-        # Live Scrollback Preview Pane (Bottom Half)
-        preview_y = start_y + split_h + 1
-        self.safe_addstr(preview_y - 1, 2, "─" * (w - 4), self._attr("dim"))
-
-        sel_pane = panes[self.sel_pane_idx] if (panes and 0 <= self.sel_pane_idx < len(panes)) else None
-        if sel_pane:
-            prev_hdr = f"LIVE SCROLLBACK PREVIEW — {sel_pane.session}:{sel_pane.pane_id} ({sel_pane.current_command}) on {sel_pane.socket}"
-            self.safe_addstr(preview_y, 2, prev_hdr, self._attr("yellow"))
-
-            txt = capture_pane_text(sel_pane.socket, sel_pane.pane_id, lines=h - preview_y - 4)
-            p_lines = txt.strip().splitlines()
-            for r_i, l_str in enumerate(p_lines[:h - preview_y - 4]):
-                self.safe_addstr(preview_y + 1 + r_i, 3, l_str, self._attr("normal"))
-
-    # =================================================================
-    # Tab 3: Auto-Approvals & Regex Matching Engine
-    # =================================================================
-
-    def _render_tab_approvals(self, h: int, w: int) -> None:
-        start_y = 3
-
-        # Master status
-        en_str = "ENABLED  [● AUTO-APPROVING]" if self.approver_state.global_enabled else "DISABLED [○ MANUAL APPROVALS ONLY]"
-        en_attr = self._attr("green") if self.approver_state.global_enabled else self._attr("warn_banner")
-        self.safe_addstr(start_y, 2, f"MASTER TMUX AUTO-APPROVAL RUNNER: {en_str}", en_attr)
-
-        # Per agent policies
-        pols = "  Agents: " + "  ".join([f"{a}: {'ON [✔]' if self.approver_state.agents_enabled.get(a, True) else 'OFF [✖]'}" for a in FLEET_AGENTS])
-        self.safe_addstr(start_y + 1, 2, pols, self._attr("dim"))
-        self.safe_addstr(start_y + 2, 2, "─" * (w - 4), self._attr("dim"))
-
-        # Rules Table
-        self.safe_addstr(start_y + 3, 2, "ACTIVE TERMINAL REGEX APPROVAL RULES:", self._attr("cyan"))
-        hdr = f"  {'STATUS':<8} {'RULE ID':<26} {'CATEGORY':<14} {'KEY':<6} {'DESCRIPTION'}"
-        self.safe_addstr(start_y + 4, 2, hdr, self._attr("bold"))
-        self.safe_addstr(start_y + 5, 2, "─" * (w - 4), self._attr("dim"))
-
-        rules = self.approver_state.rules
-        for idx, r in enumerate(rules[:6]):
-            row_y = start_y + 6 + idx
-            is_sel = (idx == self.sel_rule_idx)
-            st_tag = "ACTIVE" if r.get("enabled") else "OFF"
-            line = f"  {st_tag:<8} {r.get('id'):<26} {r.get('category'):<14} {r.get('response_key'):<6} {r.get('description', '')[:35]}"
-            attr = self._attr("selected") if is_sel else self._attr("normal")
-            self.safe_addstr(row_y, 2, " " * (w - 4), attr if is_sel else 0)
-            self.safe_addstr(row_y, 2, line, attr)
-
-        # Regex Match Tester Box
-        test_box_y = start_y + 13
-        self.safe_addstr(test_box_y, 2, "─" * (w - 4), self._attr("dim"))
-        self.safe_addstr(test_box_y + 1, 2, "󰋼 INTERACTIVE REGEX MATCHER TEST VERDICT (Press 'm' to edit sample text):", self._attr("yellow"))
-
-        if self.test_match_verdict:
-            matched = self.test_match_verdict.get("matched")
-            if matched:
-                r_name = self.test_match_verdict.get("rule_name")
-                k = self.test_match_verdict.get("key")
-                res_str = f"✔ MATCHED: Rule '{r_name}' -> Auto-Replies: '{k}'"
-                self.safe_addstr(test_box_y + 2, 4, res_str, self._attr("green"))
-            elif self.test_match_verdict.get("is_blocked"):
-                b_reason = self.test_match_verdict.get("blocked_reason")
-                self.safe_addstr(test_box_y + 2, 4, f"✖ BLOCKED: {b_reason}", self._attr("red"))
-            else:
-                self.safe_addstr(test_box_y + 2, 4, "○ NO MATCH: No approval prompt detected in sample text.", self._attr("dim"))
-
-            # Excerpt
-            exc = self.test_match_verdict.get("excerpt") or ""
-            if exc:
-                self.safe_addstr(test_box_y + 3, 4, f"Matched Excerpt: '{exc.strip().replace(chr(10), ' ')[:70]}'", self._attr("cyan"))
-
-    # =================================================================
-    # Tab 4: Surface & Logs (https://box.muse-dev.online/)
-    # =================================================================
-
-    def _render_tab_logs(self, h: int, w: int) -> None:
-        start_y = 3
-        self.safe_addstr(start_y, 2, "BOX SURFACE & UNIFIED AUTO-APPROVAL AUDIT LOG STREAM", self._attr("cyan"))
-        self.safe_addstr(start_y + 1, 2, "Surface link: https://box.muse-dev.online/  ·  HTTPS Exec: https://exec.muse-dev.online/exec", self._attr("dim"))
-        self.safe_addstr(start_y + 2, 2, "─" * (w - 4), self._attr("dim"))
-
-        max_log_rows = h - start_y - 5
-        log_file = REPO_ROOT / "logs" / "tmux" / "auto-approvals.jsonl"
-
-        lines = []
-        if log_file.exists():
-            try:
-                with open(log_file) as f:
-                    lines = f.readlines()
-            except Exception:
-                pass
-
-        if not lines:
-            self.safe_addstr(start_y + 4, 4, "(No auto-approval log events recorded yet. Press 'o' on Rules tab to run once)", self._attr("dim"))
-            return
-
-        tail = lines[-(max_log_rows + self.sel_log_scroll):]
-        if self.sel_log_scroll > 0:
-            tail = tail[:-self.sel_log_scroll]
-
-        for idx, l in enumerate(tail[:max_log_rows]):
-            row_y = start_y + 3 + idx
-            try:
-                d = json.loads(l)
-                ts = d.get("timestamp", "")[:19].replace("T", " ")
-                act = d.get("action", "")
-                ag = d.get("agent", "")
-                sess = d.get("session", "")
-                pane = d.get("pane", "")
-                key = d.get("key_sent", "")
-                rule = d.get("rule_name", "")
-                line_str = f"  [{ts}] {act:<14} {ag.upper():<6} {sess:<12} ({pane}) -> sent '{key}' [{rule}]"
-                attr = self._attr("green") if act == "AUTO_APPROVED" else (self._attr("cyan") if "DRY" in act else self._attr("warn_banner"))
-            except Exception:
-                line_str = f"  {l.strip()}"
-                attr = self._attr("normal")
-            self.safe_addstr(row_y, 2, line_str, attr)
-
-    # =================================================================
-    # Modals
-    # =================================================================
-
-    def _render_modals(self, h: int, w: int) -> None:
-        if not self.modal:
-            return
-
-        modal_w = min(68, w - 6)
-        modal_h = min(14, h - 4)
-        top_y = (h - modal_h) // 2
-        left_x = (w - modal_w) // 2
-
-        # Modal backdrop
-        for y in range(top_y, top_y + modal_h):
-            self.safe_addstr(y, left_x, " " * modal_w, self._attr("selected"))
-
-        # Border
-        self.safe_addstr(top_y, left_x, "┌" + "─" * (modal_w - 2) + "┐", self._attr("cyan"))
-        for y in range(top_y + 1, top_y + modal_h - 1):
-            self.safe_addstr(y, left_x, "│", self._attr("cyan"))
-            self.safe_addstr(y, left_x + modal_w - 1, "│", self._attr("cyan"))
-        self.safe_addstr(top_y + modal_h - 1, left_x, "└" + "─" * (modal_w - 2) + "┘", self._attr("cyan"))
-
-        if self.modal == "help":
-            self.safe_addstr(top_y + 1, left_x + 3, "󰋼 KEYBOARD CHEAT-SHEET", self._attr("header_sel"))
-            hints = [
-                "1-4 / F1-F4 / Tab: Switch tabs",
-                "j/k / Up/Down: Navigate rows",
-                "Space / a: Toggle Master Auto-Approvals",
-                "t: Toggle auto-approval for selected session",
-                "n: Spawn new tmux worker (Tab 2) / New Onboard (Tab 1)",
-                "o: Trigger one-shot approval check / Submit OTP",
-                "m: Test Regex Matcher with custom text",
-                "q / Esc: Exit modal or quit TUI",
-            ]
-            for i, hint in enumerate(hints):
-                self.safe_addstr(top_y + 3 + i, left_x + 4, hint, self._attr("normal"))
-
-        elif self.modal == "spawn_worker":
-            self.safe_addstr(top_y + 1, left_x + 3, "SPAWN NEW TMUX WORKER", self._attr("header_sel"))
-            self.safe_addstr(top_y + 3, left_x + 3, "Enter session name & command:", self._attr("bold"))
-            self.safe_addstr(top_y + 5, left_x + 3, f"> {self.modal_input_buf}_", self._attr("cyan"))
-            self.safe_addstr(top_y + 7, left_x + 3, "Format: <session_name> [command] (e.g. dev-runner python3 worker.py)", self._attr("dim"))
-            self.safe_addstr(top_y + modal_h - 2, left_x + 3, "[Enter] Spawn  [Esc] Cancel", self._attr("dim"))
-
-        elif self.modal == "test_regex":
-            self.safe_addstr(top_y + 1, left_x + 3, "EDIT TEST SAMPLE FOR REGEX MATCHER", self._attr("header_sel"))
-            self.safe_addstr(top_y + 3, left_x + 3, "Enter prompt excerpt to test:", self._attr("bold"))
-            self.safe_addstr(top_y + 5, left_x + 3, f"> {self.modal_input_buf[:55]}_", self._attr("cyan"))
-            self.safe_addstr(top_y + modal_h - 2, left_x + 3, "[Enter] Evaluate Match  [Esc] Cancel", self._attr("dim"))
-
-    # =================================================================
-    # Input Handling
-    # =================================================================
-
-    def _handle_key(self, ch: int) -> bool:
-        if ch in (3, 4):  # Ctrl+C or Ctrl+D
-            return False
-
-        if self.modal:
-            if ch in (27,):  # Esc
-                self.modal = None
-                return True
-            if self.modal in ("spawn_worker", "test_regex"):
-                if ch in (curses.KEY_ENTER, 10, 13):
-                    if self.modal == "spawn_worker":
-                        parts = self.modal_input_buf.strip().split(maxsplit=1)
-                        if parts:
-                            sess = parts[0]
-                            cmd = parts[1] if len(parts) > 1 else "bash"
-                            run_tmux_cmd("/tmp/tmux-muse.sock", "new-session", "-d", "-s", sess, cmd)
-                            self._refresh_tally()
-                            self.set_toast(f"✔ Spawned worker '{sess}' running '{cmd}'", "success")
-                        self.modal = None
-                    elif self.modal == "test_regex":
-                        self.test_text_buf = self.modal_input_buf
-                        self._eval_test_match()
-                        self.set_toast("Evaluated regex test text", "info")
-                        self.modal = None
-                    return True
-                elif ch in (curses.KEY_BACKSPACE, 127, 8):
-                    self.modal_input_buf = self.modal_input_buf[:-1]
-                    return True
-                elif 32 <= ch <= 126:
-                    self.modal_input_buf += chr(ch)
-                    return True
-            elif self.modal == "help":
-                self.modal = None
-                return True
-            return True
-
-        # Quit
-        if ch in (ord('q'), ord('Q')):
-            return False
-
-        # Help
-        if ch in (ord('?'), curses.KEY_F1):
-            self.modal = "help"
-            return True
-
-        # Tabs navigation: 1-4, F1-F4, Tab
-        if ch in (ord('1'),):
-            self.current_tab = 0
-            return True
-        elif ch in (ord('2'),):
-            self.current_tab = 1
-            return True
-        elif ch in (ord('3'),):
-            self.current_tab = 2
-            return True
-        elif ch in (ord('4'),):
-            self.current_tab = 3
-            return True
-        elif ch in (ord('\t'),):  # Tab
-            self.current_tab = (self.current_tab + 1) % len(self.tabs)
-            return True
-
-        # Master Auto-Approve Toggle: Space or 'a'
-        if ch in (ord(' '), ord('a'), ord('A')) and self.current_tab in (1, 2):
-            self.approver_state.global_enabled = not self.approver_state.global_enabled
-            self.approver_state.save()
-            state_str = "ENABLED" if self.approver_state.global_enabled else "DISABLED"
-            self.set_toast(f"Master Auto-Approvals: {state_str}", "success" if self.approver_state.global_enabled else "warn")
-            self._refresh_tally()
-            return True
-
-        # Tab 0: Onboard Connects keys
-        if self.current_tab == 0:
-            if ch in (ord('j'), curses.KEY_DOWN):
-                self.sel_connect_idx = min(len(self.connects) - 1, self.sel_connect_idx + 1)
-                return True
-            elif ch in (ord('k'), curses.KEY_UP):
-                self.sel_connect_idx = max(0, self.sel_connect_idx - 1)
-                return True
-            elif ch in (ord('r'), ord('R')):
-                self._refresh_connects()
-                self.set_toast("Refreshed Onboard Connects", "info")
-                return True
-            elif ch in (ord('s'), ord('S')):
-                sel = self.connects[self.sel_connect_idx] if 0 <= self.sel_connect_idx < len(self.connects) else None
-                node = sel.get("node", "646") if sel else "646"
-                self.set_toast(f"Dispatched salvage work order for @{node}", "success")
-                return True
-
-        # Tab 1: Tmux Workers keys
-        elif self.current_tab == 1:
-            panes = self.tally.panes
-            if ch in (ord('j'), curses.KEY_DOWN):
-                self.sel_pane_idx = min(len(panes) - 1, self.sel_pane_idx + 1)
-                return True
-            elif ch in (ord('k'), curses.KEY_UP):
-                self.sel_pane_idx = max(0, self.sel_pane_idx - 1)
-                return True
-            elif ch in (ord('t'), ord('T')):
-                if panes and 0 <= self.sel_pane_idx < len(panes):
-                    p = panes[self.sel_pane_idx]
-                    cur = self.approver_state.sessions_enabled.get(p.session, True)
-                    self.approver_state.sessions_enabled[p.session] = not cur
-                    self.approver_state.save()
-                    self._refresh_tally()
-                    state_str = "ON" if not cur else "OFF"
-                    self.set_toast(f"Toggled Auto-Approve for '{p.session}': {state_str}", "info")
-                return True
-            elif ch in (ord('n'), ord('N')):
-                self.modal = "spawn_worker"
-                self.modal_input_buf = ""
-                return True
-            elif ch in (ord('k'), ord('K')):
-                if panes and 0 <= self.sel_pane_idx < len(panes):
-                    p = panes[self.sel_pane_idx]
-                    run_tmux_cmd(p.socket, "kill-session", "-t", p.session)
-                    self._refresh_tally()
-                    self.set_toast(f"✔ Killed session '{p.session}'", "warn")
-                return True
-            elif ch in (ord('r'), ord('R')):
-                self._refresh_tally()
-                self.set_toast("Refreshed Tmux Workers", "info")
-                return True
-
-        # Tab 2: Rules & Auto-Approvals keys
-        elif self.current_tab == 2:
-            if ch in (ord('m'), ord('M')):
-                self.modal = "test_regex"
-                self.modal_input_buf = self.test_text_buf
-                return True
-            elif ch in (ord('o'), ord('O')):
-                res = self.runner.run_once()
-                self.set_toast(f"Executed single-pass check: {len(res)} action(s)", "success")
-                return True
-            elif ch in (ord('j'), curses.KEY_DOWN):
-                self.sel_rule_idx = min(len(self.approver_state.rules) - 1, self.sel_rule_idx + 1)
-                return True
-            elif ch in (ord('k'), curses.KEY_UP):
-                self.sel_rule_idx = max(0, self.sel_rule_idx - 1)
-                return True
-
-        # Tab 3: Logs keys
-        elif self.current_tab == 3:
-            if ch in (ord('j'), curses.KEY_DOWN):
-                self.sel_log_scroll = max(0, self.sel_log_scroll - 1)
-                return True
-            elif ch in (ord('k'), curses.KEY_UP):
-                self.sel_log_scroll += 1
-                return True
-            elif ch in (ord('e'), ord('E')):
-                self.set_toast("Synced status with https://box.muse-dev.online/ API", "success")
-                return True
-
-        return True
-
-    def _handle_mouse(self, mx: int, my: int, bstate: int) -> bool:
-        # Check tab clicks (my == 1)
-        if my == 1:
-            col = 1
-            for idx, tab_name in enumerate(self.tabs):
-                tab_w = len(tab_name) + 4
-                if col <= mx < col + tab_w:
-                    self.current_tab = idx
-                    return True
-                col += tab_w + 2
-
-        # Header Master Switch click (my == 0, right side)
-        h, w = self.stdscr.getmaxyx()
-        if my == 0 and mx >= w - 30:
-            self.approver_state.global_enabled = not self.approver_state.global_enabled
-            self.approver_state.save()
-            self._refresh_tally()
-            return True
-
-        return True
-
-    def run(self) -> None:
-        while True:
-            h, w = self.stdscr.getmaxyx()
-            self.stdscr.erase()
-
-            self._render_header(w)
-
-            if self.current_tab == 0:
-                self._render_tab_onboard(h, w)
-            elif self.current_tab == 1:
-                self._render_tab_tmux(h, w)
-            elif self.current_tab == 2:
-                self._render_tab_approvals(h, w)
-            else:
-                self._render_tab_logs(h, w)
-
-            self._render_footer(h, w)
-            self._render_modals(h, w)
-
-            self.stdscr.refresh()
-
-            try:
-                ch = self.stdscr.getch()
-                if ch != -1:
-                    if ch == curses.KEY_MOUSE:
-                        try:
-                            _, mx, my, _, bstate = curses.getmouse()
-                            self._handle_mouse(mx, my, bstate)
-                        except Exception:
-                            pass
-                    else:
-                        if not self._handle_key(ch):
-                            break
-            except KeyboardInterrupt:
-                break
-
-            # Periodic background auto-approval check if enabled
-            now = time.time()
-            if self.approver_state.global_enabled and now - self.last_auto_poll > 2.0:
-                self.last_auto_poll = now
-                self.runner.run_once()
-                self._refresh_tally()
-
-            time.sleep(0.04)
-
-
-def main() -> int:
-    try:
-        curses.wrapper(lambda stdscr: BoxOnboardTUI(stdscr).run())
-    finally:
-        try:
-            sys.stdout.write("\033[?1000l\033[?1002l\033[?1006l\033[?2004l")
-            sys.stdout.flush()
-        except Exception:
-            pass
-    return 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/box-query.py

@@ -1,74 +0,0 @@
-#!/usr/bin/env python3
-"""
-box-query.py — Lightweight client to query box.muse-dev.online APIs over HTTPS.
-
-Works inside containers and nodes without direct SSH access:
-Uses SSH signature authentication (?identity=bl&ts=...&sig=...) against
-the VM Box API.
-
-Usage:
-  box-query.py timers
-  box-query.py jobs
-  box-query.py agents
-  box-query.py dms [limit]
-"""
-
-import argparse
-import json
-import os
-import subprocess
-import sys
-import time
-import urllib.parse
-import urllib.request
-
-BOX_API_BASE = os.environ.get("BOX_API_BASE", "https://box.muse-dev.online")
-BOX_SIGN_KEY = os.environ.get("BOX_SIGN_KEY", os.path.expanduser("~/.ssh/id_ed25519"))
-
-def sign_request(identity, endpoint):
-    ts = str(int(time.time()))
-    payload = f"{ts}\n{endpoint}".encode()
-    try:
-        p = subprocess.run(
-            ["ssh-keygen", "-Y", "sign", "-f", BOX_SIGN_KEY, "-n", "box"],
-            input=payload, capture_output=True, timeout=15)
-        if p.returncode != 0:
-            return None, None
-        return ts, p.stdout.decode()
-    except Exception:
-        return None, None
-
-def query(endpoint):
-    ts, sig = sign_request("bl", endpoint)
-    if not ts or not sig:
-        sys.stderr.write("Failed to sign request (missing key or ssh-keygen error)\n")
-        sys.exit(1)
-
-    query_str = urllib.parse.urlencode({
-        "identity": "bl",
-        "ts": ts,
-        "sig": sig
-    })
-    url = f"{BOX_API_BASE}/api/box/{endpoint}?{query_str}"
-    req = urllib.request.Request(
-        url,
-        headers={"User-Agent": "NetVM-box-query/1.0 (container)"}
-    )
-    with urllib.request.urlopen(req, timeout=15) as resp:
-        data = resp.read().decode("utf-8")
-        try:
-            return json.loads(data)
-        except Exception:
-            return data
-
-def main():
-    p = argparse.ArgumentParser(description="Query box.muse-dev.online APIs via HTTPS signature auth")
-    p.add_argument("endpoint", choices=["timers", "jobs", "agents", "dms", "health"])
-    p.add_argument("--json", action="store_true")
-    args = p.parse_args()
-
-    res = query(args.endpoint)
-    print(json.dumps(res, indent=2))
-
-if __name__ == "__main__":
-    main()

bin/box-readback-loopback.py

@@ -0,0 +1,424 @@
+#!/usr/bin/env python3
+"""box-readback-loopback.py — Synchronous Readback Gate & Cognitive-Aware True Loopback Engine.
+
+Architecture:
+1. Readback Gate:
+   - Synchronously awaits agent confirmation (up to 45s) after task dispatch.
+   - Validates via Hybrid Tag ([READBACK] Ticket #...) + Semantic Fallback.
+   - Marks ticket 'in-progress' upon confirmation; marks 'blocked' and releases agent on timeout.
+2. Cognitive-Aware True Loopbacks:
+   - Zero Token Burn: 100% silent while git commits or PRs are progressing.
+   - Cognitive Guard: Defer loopbacks while agent is THINKING / GENERATING.
+   - Dual-Layer Escalation:
+     * 15m inactive: Tier 1 non-intrusive Gitea ticket comment (@agent inquiry).
+     * 45m inactive: Tier 2 direct chat DM escalation (muse-cli-node send).
+     * 90m inactive: Tier 3 failure escalation (mark 'blocked', alert #lobby, release agent).
+"""
+
+import json
+import os
+import re
+import socket
+import subprocess
+import sys
+import time
+import urllib.request
+from datetime import datetime, timezone
+from pathlib import Path
+
+# Local imports
+try:
+    import agent_cognitive_probe as acp
+except ImportError:
+    acp = None
+
+REMOTE_HOST = "100.123.153.75"  # bl control node
+DEFAULT_GITEA_URL = "https://tea.muse-dev.online"
+LOOPBACK_STATE_FILE = Path("/tmp/box-loopback-state.json")
+
+
+def is_running_on_bl():
+    try:
+        hn = socket.gethostname().lower()
+        if "bl" in hn:
+            return True
+    except Exception:
+        pass
+    return os.path.exists("/var/run/netns/warp-muse") or os.path.exists("/run/netns/warp-muse")
+
+
+def get_gitea_token():
+    token = os.environ.get("GITEA_TOKEN", "3c26744525bceaf385aa09737f7e41af613627b6")
+    return token
+
+
+def gitea_api(endpoint: str, method: str = "GET", data: dict = None):
+    token = get_gitea_token()
+    url = f"{DEFAULT_GITEA_URL}/api/v1{endpoint}"
+    headers = {
+        "Authorization": f"token {token}",
+        "Content-Type": "application/json",
+        "User-Agent": "Box-Work-CLI/1.0",
+    }
+    payload = json.dumps(data).encode("utf-8") if data else None
+    req = urllib.request.Request(url, data=payload, headers=headers, method=method)
+    try:
+        with urllib.request.urlopen(req, timeout=10) as r:
+            if r.status in (200, 201):
+                return json.loads(r.read().decode())
+            return {"status": r.status}
+    except urllib.error.HTTPError as e:
+        try:
+            return json.loads(e.read().decode())
+        except Exception:
+            return {"error": str(e), "code": e.code}
+    except Exception as e:
+        return {"error": str(e)}
+
+
+# ----------------------------------------------------------------------
+# CHAT COMMUNICATION HELPERS
+# ----------------------------------------------------------------------
+
+def send_agent_chat(agent: str, message: str) -> bool:
+    """Delivers a message directly into the agent's web chat session."""
+    if is_running_on_bl():
+        cmd = ["/home/super/Projects/NetVM/bin/muse-cli-node", agent, "send", message]
+    else:
+        cmd = ["ssh", "-q", f"super@{REMOTE_HOST}",
+               f"/home/super/Projects/NetVM/bin/muse-cli-node {agent} send {subprocess.list2cmdline([message])}"]
+    try:
+        res = subprocess.run(cmd, capture_output=True, text=True, timeout=15)
+        return res.returncode == 0
+    except Exception:
+        return False
+
+
+def get_agent_history(agent: str, limit: int = 5) -> list:
+    """Retrieves recent chat messages from the agent's active session."""
+    if is_running_on_bl():
+        cmd = ["/home/super/Projects/NetVM/bin/muse-cli-node", agent, "history", "--limit", str(limit)]
+    else:
+        cmd = ["ssh", "-q", f"super@{REMOTE_HOST}",
+               f"/home/super/Projects/NetVM/bin/muse-cli-node {agent} history --limit {limit}"]
+    try:
+        res = subprocess.run(cmd, capture_output=True, text=True, timeout=15)
+        if res.returncode == 0 and res.stdout.strip():
+            data = json.loads(res.stdout.strip())
+            if isinstance(data, list):
+                return data
+    except Exception:
+        pass
+    return []
+
+
+def get_latest_chat_seq(agent: str) -> int:
+    """Finds the maximum sequence number in the agent's chat history."""
+    history = get_agent_history(agent, limit=3)
+    seqs = [m.get("seq", 0) for m in history if isinstance(m, dict) and "seq" in m]
+    return max(seqs) if seqs else 0
+
+
+# ----------------------------------------------------------------------
+# READBACK GATE
+# ----------------------------------------------------------------------
+
+def validate_readback(text: str, issue_num: int, agent: str) -> tuple[bool, str]:
+    """Validates an agent readback using Hybrid Tag + Semantic Fallback.
+    
+    Returns (is_valid, excerpt).
+    """
+    if not text:
+        return False, ""
+
+    clean_text = text.strip()
+    issue_pattern = rf"#?{issue_num}\b"
+
+    # 1. Strict Tag Match: [READBACK] Ticket #<num> ...
+    if re.search(r"\[READBACK\]", clean_text, re.IGNORECASE) and re.search(issue_pattern, clean_text):
+        snippet = clean_text[:200].replace("\n", " ")
+        return True, snippet
+
+    # 2. Semantic Fallback: Mentions ticket number AND branch/accepted status
+    has_issue = bool(re.search(issue_pattern, clean_text))
+    has_branch_or_ack = bool(re.search(
+        rf"(dev/{agent}/|branch|accepted|working on|confirm|start(ed|ing)|received)",
+        clean_text, re.IGNORECASE
+    ))
+
+    if has_issue and has_branch_or_ack:
+        snippet = clean_text[:200].replace("\n", " ")
+        return True, snippet
+
+    return False, ""
+
+
+def wait_for_readback(agent: str, issue_num: int, initial_seq: int, timeout_s: int = 45, poll_s: float = 3.0) -> dict:
+    """Synchronously polls for agent readback within timeout_s."""
+    start_time = time.time()
+    deadline = start_time + timeout_s
+
+    while time.time() < deadline:
+        elapsed = int(time.time() - start_time)
+        print(f"\r  ⏳ Awaiting Readback from @{agent} ({elapsed}s / {timeout_s}s)...", end="", flush=True)
+
+        history = get_agent_history(agent, limit=4)
+        for msg in history:
+            seq = msg.get("seq", 0)
+            role = msg.get("role", "")
+            text = msg.get("text", "")
+
+            # Only check new assistant messages
+            if seq > initial_seq and role == "assistant":
+                valid, excerpt = validate_readback(text, issue_num, agent)
+                if valid:
+                    print()
+                    return {
+                        "success": True,
+                        "snippet": excerpt,
+                        "elapsed": elapsed,
+                        "seq": seq,
+                    }
+
+        time.sleep(poll_s)
+
+    print()
+    return {
+        "success": False,
+        "timeout": True,
+        "elapsed": timeout_s,
+    }
+
+
+def handle_readback_success(agent: str, issue_num: int, snippet: str):
+    """Marks ticket in-progress and records confirmation on Gitea."""
+    # Label ticket in-progress
+    gitea_api(f"/repos/super/box/issues/{issue_num}/labels", method="POST", data={"labels": ["in-progress"]})
+    # Post confirmation comment
+    comment_body = f"🤖 **Readback Confirmed** by @{agent}:\n> {snippet}"
+    gitea_api(f"/repos/super/box/issues/{issue_num}/comments", method="POST", data={"body": comment_body})
+
+
+def handle_readback_timeout(agent: str, issue_num: int, title: str):
+    """Labels ticket blocked and unassigns agent so they return to IDLE."""
+    # Label ticket blocked
+    gitea_api(f"/repos/super/box/issues/{issue_num}/labels", method="POST", data={"labels": ["blocked"]})
+    # Post explanation comment
+    comment_body = (
+        f"⚠️ **Readback Timeout**: Agent @{agent} did not confirm ticket #{issue_num} "
+        f"within 45 seconds of dispatch. Releasing assignment to prevent deadlocks."
+    )
+    gitea_api(f"/repos/super/box/issues/{issue_num}/comments", method="POST", data={"body": comment_body})
+    # Unassign agent
+    gitea_api(f"/repos/super/box/issues/{issue_num}", method="PATCH", data={"assignees": []})
+
+
+# ----------------------------------------------------------------------
+# COGNITIVE-AWARE TRUE LOOPBACK ENGINE
+# ----------------------------------------------------------------------
+
+def load_loopback_state() -> dict:
+    if LOOPBACK_STATE_FILE.exists():
+        try:
+            with open(LOOPBACK_STATE_FILE) as f:
+                return json.load(f)
+        except Exception:
+            pass
+    return {}
+
+
+def save_loopback_state(state: dict):
+    try:
+        with open(LOOPBACK_STATE_FILE, "w") as f:
+            json.dump(state, f, indent=2)
+    except Exception:
+        pass
+
+
+def get_ticket_git_activity(agent: str, issue_num: int) -> datetime | None:
+    """Checks the latest commit timestamp on the agent's branch dev/<agent>/<issue>-*."""
+    # Check Gitea branches for dev/<agent>/<issue_num>-*
+    branches = gitea_api("/repos/super/box/branches")
+    if isinstance(branches, list):
+        target_prefix = f"dev/{agent}/{issue_num}"
+        for b in branches:
+            name = b.get("name", "")
+            if target_prefix in name:
+                commit = b.get("commit", {})
+                ts_str = commit.get("timestamp")
+                if ts_str:
+                    try:
+                        return datetime.fromisoformat(ts_str.replace("Z", "+00:00"))
+                    except Exception:
+                        pass
+    return None
+
+
+def get_ticket_last_activity(issue: dict, agent: str) -> tuple[datetime, str]:
+    """Finds the most recent activity timestamp (git commit, comment, or issue creation)."""
+    issue_num = issue["number"]
+    latest_dt = datetime.fromisoformat(issue["created_at"].replace("Z", "+00:00"))
+    source = "issue_created"
+
+    # Check comments
+    comments = gitea_api(f"/repos/super/box/issues/{issue_num}/comments")
+    if isinstance(comments, list):
+        for c in comments:
+            c_dt = datetime.fromisoformat(c["created_at"].replace("Z", "+00:00"))
+            if c_dt > latest_dt:
+                latest_dt = c_dt
+                source = "gitea_comment"
+
+    # Check git branch commit
+    git_dt = get_ticket_git_activity(agent, issue_num)
+    if git_dt and git_dt > latest_dt:
+        latest_dt = git_dt
+        source = "git_commit"
+
+    return latest_dt, source
+
+
+def run_loopback_sweep(dry_run: bool = False, verbose: bool = True) -> list:
+    """Executes a single sweep of all open assigned tickets according to the 3-tier escalation model."""
+    now = datetime.now(timezone.utc)
+    state = load_loopback_state()
+    actions_taken = []
+
+    issues = gitea_api("/repos/super/box/issues?state=open")
+    if not isinstance(issues, list):
+        if verbose:
+            print("Failed to fetch open issues from Gitea.")
+        return []
+
+    assigned_issues = [i for i in issues if i.get("assignee")]
+    if verbose:
+        print(f"\n=== LOOPBACK SWEEP: {len(assigned_issues)} ACTIVE ASSIGNED TICKETS ({now.strftime('%H:%M:%SZ')}) ===")
+
+    for iss in assigned_issues:
+        issue_num = iss["number"]
+        title = iss.get("title", "")
+        agent = iss["assignee"]["username"]
+        key = str(issue_num)
+        ticket_state = state.get(key, {})
+
+        last_dt, source = get_ticket_last_activity(iss, agent)
+        inactive_s = (now - last_dt).total_seconds()
+        inactive_m = int(inactive_s // 60)
+
+        # Check cognitive state
+        cog = acp.get_passive_cognitive_state(agent) if acp else {"status": "IDLE", "cognitive_lock": False}
+        cog_status = cog.get("status", "IDLE")
+        is_thinking = cog_status == "THINKING" or cog.get("cognitive_lock")
+
+        if verbose:
+            print(f"Ticket #{issue_num} (@{agent}): {inactive_m}m inactive (source: {source}) | Cognitive: {cog_status}")
+
+        # Tier 0: Inactive < 15m or active git commits -> Complete silence
+        if inactive_m < 15 or source == "git_commit":
+            if verbose:
+                print("  👉 Status: Active or within silent grace period (<15m). No action.")
+            continue
+
+        # Check cognitive guard: Defer if agent is thinking/generating
+        if is_thinking and cog_status != "INPUT_WAIT":
+            if verbose:
+                print(f"  🧠 Cognitive Guard: Deferring loopback — @{agent} is currently {cog_status}.")
+            continue
+
+        # Tier 1: 15m <= Inactivity < 45m -> Non-intrusive Gitea ticket comment
+        if 15 <= inactive_m < 45:
+            if ticket_state.get("tier1_sent"):
+                if verbose:
+                    print("  👉 Tier 1 comment already dispatched. Waiting for 45m threshold.")
+                continue
+
+            msg = (
+                f"🤖 @{agent} **Loopback Tier 1 Check** ({inactive_m}m elapsed):\n"
+                f"No git commits recorded on feature branch for Ticket #{issue_num}. "
+                f"Are you progressing or blocked? Reply with status or push a commit."
+            )
+            action_desc = f"Tier 1: Posted Gitea comment to #{issue_num} (@{agent})"
+            actions_taken.append(action_desc)
+
+            if not dry_run:
+                gitea_api(f"/repos/super/box/issues/{issue_num}/comments", method="POST", data={"body": msg})
+                ticket_state["tier1_sent"] = now.isoformat()
+                state[key] = ticket_state
+                save_loopback_state(state)
+            if verbose:
+                print(f"  ✓ {action_desc}")
+
+        # Tier 2: 45m <= Inactivity < 90m -> Direct Chat DM Escalation
+        elif 45 <= inactive_m < 90:
+            if ticket_state.get("tier2_sent"):
+                if verbose:
+                    print("  👉 Tier 2 chat DM already dispatched. Waiting for 90m threshold.")
+                continue
+
+            chat_msg = (
+                f"[LOOPBACK ALERT] Ticket #{issue_num} ('{title}'): "
+                f"{inactive_m} minutes inactive with no git commits. "
+                f"Please confirm if blocked on tool execution, terminal approvals, or environment."
+            )
+            action_desc = f"Tier 2: Escalated to chat DM for @{agent} on #{issue_num}"
+            actions_taken.append(action_desc)
+
+            if not dry_run:
+                send_agent_chat(agent, chat_msg)
+                gitea_api(f"/repos/super/box/issues/{issue_num}/comments", method="POST", data={
+                    "body": f"📣 **Loopback Tier 2 Escalation**: Inactivity reached {inactive_m}m. Sent direct chat DM to @{agent}."
+                })
+                ticket_state["tier2_sent"] = now.isoformat()
+                state[key] = ticket_state
+                save_loopback_state(state)
+            if verbose:
+                print(f"  ✓ {action_desc}")
+
+        # Tier 3: Inactivity >= 90m (or unhandled INPUT_WAIT > 15m) -> Fail-closed escalation
+        elif inactive_m >= 90 or (cog_status == "INPUT_WAIT" and inactive_m >= 15):
+            action_desc = f"Tier 3: Ticket #{issue_num} marked BLOCKED; released @{agent} assignment"
+            actions_taken.append(action_desc)
+
+            if not dry_run:
+                # Label blocked
+                gitea_api(f"/repos/super/box/issues/{issue_num}/labels", method="POST", data={"labels": ["blocked"]})
+                # Post failure comment
+                gitea_api(f"/repos/super/box/issues/{issue_num}/comments", method="POST", data={
+                    "body": (
+                        f"🚨 **Loopback Tier 3 Escalation**: Inactivity reached {inactive_m}m with zero git progress. "
+                        f"Ticket marked `blocked` and unassigned from @{agent} for operator intervention."
+                    )
+                })
+                # Unassign agent
+                gitea_api(f"/repos/super/box/issues/{issue_num}", method="PATCH", data={"assignees": []})
+                ticket_state["tier3_sent"] = now.isoformat()
+                state[key] = ticket_state
+                save_loopback_state(state)
+            if verbose:
+                print(f"  🚨 {action_desc}")
+
+    if verbose:
+        print()
+    return actions_taken
+
+
+def main():
+    import argparse
+    parser = argparse.ArgumentParser(description="Synchronous Readback & Cognitive True Loopback Engine")
+    sub = parser.add_subparsers(dest="cmd")
+
+    p_sweep = sub.add_parser("sweep", help="Run a loopback sweep across open tickets")
+    p_sweep.add_argument("--dry-run", action="store_true", help="Evaluate conditions without sending messages")
+    p_sweep.add_argument("--json", action="store_true", help="Output actions as JSON")
+
+    args = parser.parse_args()
+    if not args.cmd or args.cmd == "sweep":
+        dry_run = getattr(args, "dry_run", False)
+        as_json = getattr(args, "json", False)
+        actions = run_loopback_sweep(dry_run=dry_run, verbose=not as_json)
+        if as_json:
+            print(json.dumps({"ok": True, "actions": actions}, indent=2))
+
+
+if __name__ == "__main__":
+    main()

bin/box-stability-watcher.py

@@ -1 +0,0 @@
-../watchers/box-stability-watcher.py

bin/box-sys-op.py

@@ -1,204 +0,0 @@
-#!/usr/bin/env python3
-"""
-box-sys-op.py — Sandboxed execution helper for core system operations:
-  files.read, files.write, web.fetch, service.status, service.restart
-Called via fixed argv from exec-constrained.py.
-"""
-import sys
-import os
-import json
-import urllib.request
-import urllib.error
-import urllib.parse
-import ipaddress
-import subprocess
-from pathlib import Path
-
-REPO_ROOT = Path("/home/super/Projects/NetVM").resolve()
-MAX_OUTPUT = 4096
-ALLOWED_SERVICES = {
-    "board.service", "caddy.service", "response-harvester.timer",
-    "self-main-loop.timer", "job-heartbeat.timer", "job-scheduler.timer"
-}
-
-def safe_repo_path(raw):
-    clean = os.path.normpath(raw.strip())
-    if not os.path.isabs(clean):
-        clean = os.path.normpath(str(REPO_ROOT / clean))
-    real = Path(clean).resolve()
-    if not str(real).startswith(str(REPO_ROOT) + "/") and real != REPO_ROOT:
-        raise ValueError("path must reside inside repository root (/home/super/Projects/NetVM)")
-    return real
-
-def op_files_read(path_str, max_lines=100):
-    p = safe_repo_path(path_str)
-    if not p.exists() or not p.is_file():
-        return {"ok": False, "error": f"File not found: {path_str}"}
-    with open(p, "r", encoding="utf-8", errors="replace") as f:
-        lines = f.readlines()
-    total_lines = len(lines)
-    snippet = "".join(lines[:max_lines])
-    truncated = total_lines > max_lines or len(snippet) > MAX_OUTPUT
-    if len(snippet) > MAX_OUTPUT:
-        snippet = snippet[:MAX_OUTPUT] + "\n... [truncated]"
-    return {
-        "ok": True,
-        "path": str(p.relative_to(REPO_ROOT)),
-        "lines": total_lines,
-        "displayed_lines": min(total_lines, max_lines),
-        "content": snippet,
-        "truncated": truncated
-    }
-
-def op_files_write(path_str, content):
-    p = safe_repo_path(path_str)
-    p.parent.mkdir(parents=True, exist_ok=True)
-    with open(p, "w", encoding="utf-8") as f:
-        f.write(content)
-    return {
-        "ok": True,
-        "path": str(p.relative_to(REPO_ROOT)),
-        "bytes_written": len(content.encode("utf-8")),
-        "lines": content.count("\n") + 1
-    }
-
-def op_web_fetch(url_str):
-    parsed = urllib.parse.urlparse(url_str)
-    if parsed.scheme not in ("http", "https"):
-        return {"ok": False, "error": "URL scheme must be http or https"}
-    host = parsed.hostname or ""
-    if not host or host in ("localhost", "127.0.0.1", "::1"):
-        return {"ok": False, "error": "Loopback destinations blocked"}
-    try:
-        ip = ipaddress.ip_address(host)
-        if ip.is_private or ip.is_loopback or ip.is_link_local:
-            return {"ok": False, "error": "Private and local IP addresses blocked"}
-    except ValueError:
-        pass
-
-    req = urllib.request.Request(
-        url_str,
-        headers={"User-Agent": "Mozilla/5.0 Box-Agent-Client/1.0"}
-    )
-    try:
-        with urllib.request.urlopen(req, timeout=10) as resp:
-            data = resp.read(MAX_OUTPUT + 1024).decode("utf-8", errors="replace")
-            status = resp.status
-            truncated = len(data) > MAX_OUTPUT
-            if truncated:
-                data = data[:MAX_OUTPUT] + "\n... [truncated]"
-            return {
-                "ok": True,
-                "url": url_str,
-                "status": status,
-                "length": len(data),
-                "body": data,
-                "truncated": truncated
-            }
-    except urllib.error.HTTPError as he:
-        return {"ok": False, "status": he.code, "error": f"HTTP {he.code}: {he.reason}"}
-    except Exception as e:
-        return {"ok": False, "error": str(e)}
-
-def op_service_status(unit):
-    if unit not in ALLOWED_SERVICES:
-        return {"ok": False, "error": f"Service not allowed: {unit}"}
-    flag = "--user" if unit.endswith(".timer") else "--system"
-    cmd = ["systemctl", flag, "status", unit] if flag == "--user" else ["systemctl", "is-active", unit]
-    r = subprocess.run(cmd, capture_output=True, text=True, timeout=10)
-    active = "active" in r.stdout.lower() or "active" in r.stderr.lower()
-    return {
-        "ok": True,
-        "service": unit,
-        "active": active,
-        "status_line": r.stdout.splitlines()[0] if r.stdout.splitlines() else "unknown",
-        "output": r.stdout[:500].strip()
-    }
-
-def op_service_restart(unit):
-    if unit not in ALLOWED_SERVICES:
-        return {"ok": False, "error": f"Service not allowed: {unit}"}
-    if unit.endswith(".timer"):
-        cmd = ["systemctl", "--user", "restart", unit]
-    else:
-        cmd = ["sudo", "-n", "systemctl", "restart", unit]
-    r = subprocess.run(cmd, capture_output=True, text=True, timeout=15)
-    return {
-        "ok": r.returncode == 0,
-        "service": unit,
-        "restarted": r.returncode == 0,
-        "error": r.stderr.strip() if r.returncode != 0 else None
-    }
-
-def op_followup_schedule(args):
-    agent = args.get("agent")
-    if agent not in ("muse", "pip", "646", "opm", "dev", "def"):
-        return {"ok": False, "error": f"Invalid agent: {agent}"}
-    sender = args.get("sender") or agent
-    if sender not in ("muse", "pip", "646", "opm", "dev", "def"):
-        sender = agent
-    try:
-        in_m = float(args.get("in_m", 1))
-    except (TypeError, ValueError):
-        return {"ok": False, "error": "in_m must be a number"}
-    sec = max(5, int(in_m * 60))
-    prompt = args.get("prompt", "")
-    if not prompt or not isinstance(prompt, str):
-        return {"ok": False, "error": "prompt must be a non-empty string"}
-    if len(prompt) > 1000:
-        return {"ok": False, "error": "prompt exceeds 1000 characters"}
-
-    sidechat = args.get("thread") or args.get("sidechat")
-    cmd = [
-        "systemd-run", "--user", f"--on-active={sec}s",
-        sys.executable, str(REPO_ROOT / "bin" / "box-ctl.py"),
-        "notify", agent, prompt, "--sender", sender
-    ]
-    if sidechat and isinstance(sidechat, str) and len(sidechat) <= 64:
-        cmd.extend(["--sidechat", sidechat])
-
-    r = subprocess.run(cmd, capture_output=True, text=True, timeout=15)
-    if r.returncode != 0:
-        return {"ok": False, "error": r.stderr.strip() or r.stdout.strip()}
-    return {
-        "ok": True,
-        "agent": agent,
-        "in_seconds": sec,
-        "sidechat": sidechat,
-        "timer_info": (r.stderr or r.stdout).strip()
-    }
-
-def main():
-    if len(sys.argv) < 2:
-        print(json.dumps({"ok": False, "error": "missing operation"}))
-        sys.exit(1)
-    op = sys.argv[1]
-    raw_args = sys.stdin.read()
-    try:
-        args = json.loads(raw_args) if raw_args.strip() else {}
-    except Exception as e:
-        print(json.dumps({"ok": False, "error": f"bad json args: {e}"}))
-        sys.exit(1)
-
-    try:
-        if op == "files.read":
-            res = op_files_read(args.get("path", ""), int(args.get("lines", 100)))
-        elif op == "files.write":
-            res = op_files_write(args.get("path", ""), args.get("content", ""))
-        elif op == "web.fetch":
-            res = op_web_fetch(args.get("url", ""))
-        elif op == "service.status":
-            res = op_service_status(args.get("unit", args.get("name", "")))
-        elif op == "service.restart":
-            res = op_service_restart(args.get("unit", args.get("name", "")))
-        elif op in ("followup.create", "followup.schedule"):
-            res = op_followup_schedule(args)
-        else:
-            res = {"ok": False, "error": f"unknown operation: {op}"}
-    except Exception as e:
-        res = {"ok": False, "error": str(e)}
-
-    print(json.dumps(res))
-
-if __name__ == "__main__":
-    main()

bin/box-tui.py

@@ -1 +0,0 @@
-muse-tui.py

bin/box_readback_loopback.py

@@ -0,0 +1 @@
+box-readback-loopback.py

bin/box_work.py

@@ -0,0 +1 @@
+/home/super/Projects/NetVM/bin/box-work.py

bin/brain.py

@@ -1,413 +0,0 @@
-"""
-brain.py — Main-loop brain workspace module.
-
-The brain sidechat ("main-loop brain" on opm's account) is where the main loop
-OPERATES: it posts its thinking there, reads operator instructions from there,
-and keeps its working state visible there.
-
-Deploy to: ~/Projects/NetVM/bin/brain.py on bl (alongside self_main_loop.py).
-
-Design source: ~/workspace/main-loop-brain-design.md
-User directive 2026-10-04: "we need main loop to operate its brains in side chat"
-
-SAFETY CONTRACT (do not weaken):
-  - Brain posts carry a [BRAIN <ts>] marker, NEVER a [JOB <id>] marker.
-    The response-harvester keys off [JOB ...]; a thinking note must never
-    look actionable.
-  - Brain posts NEVER use dm.py --expect-reply. Thinking creates no followup
-    records, no nudges, no escalations.
-  - When reading the brain, the loop skips its own messages (sender check).
-    The loop must never digest its own thinking as agent activity.
-  - !loop commands are honored ONLY from AUTHORIZED_SENDERS. Everything else
-    is read as context, never as instruction.
-  - Malformed commands get a one-line correction posted to the brain.
-    Never silent, never a crash.
-  - Cap: MAX_BRAIN_POSTS_PER_TICK posts per tick. The brain must not amplify.
-
-State lives in the existing watermark JSON file under the "brain" key:
-  {"brain": {"ignores": {"646": <expires_epoch>}, "quiet_until": <epoch|0>,
-             "brain_watermark": <epoch>, "tick": <int>}}
-Absolute expiries so a dead loop cannot leave an agent ignored forever.
-"""
-
-import json
-import os
-import re
-import subprocess
-import time
-
-# ---------------------------------------------------------------------------
-# Constants
-# ---------------------------------------------------------------------------
-
-BRAIN_AGENT = "opm"
-BRAIN_SIDECHAT_NAME = "main-loop brain"
-
-# Operator identities allowed to issue !loop commands. These must match the
-# DM-signer / board identity strings; do not invent new ones here.
-AUTHORIZED_SENDERS = {"super", "operator-646", "operator-main"}
-
-# Sender identities the loop itself posts under (skipped on read-back).
-OWN_SENDERS = {"main-loop", "self_main_loop", "operator-main-loop", "opm"}
-
-BRAIN_MARKER_RE = re.compile(r"\[BRAIN\s+([^\]]+)\]")
-JOB_MARKER_RE = re.compile(r"\[JOB\s+([^\]]+)\]")
-LOOP_CMD_RE = re.compile(r"^\s*!loop\s+(\S+)(.*)$", re.IGNORECASE)
-
-MAX_BRAIN_POSTS_PER_TICK = 4
-DEFAULT_BIN_DIR = os.path.expanduser("~/Projects/NetVM/bin")
-
-VALID_AGENTS = {"muse", "pip", "646", "opm"}
-
-DUR_RE = re.compile(r"^\s*(\d+)\s*([smh])\s*$", re.IGNORECASE)
-
-
-# ---------------------------------------------------------------------------
-# Pure logic — fully unit-testable, no I/O
-# ---------------------------------------------------------------------------
-
-def parse_duration(text):
-    """'30m' -> 1800.0, '1h' -> 3600.0, '90s' -> 90.0. None if malformed."""
-    m = DUR_RE.match(text or "")
-    if not m:
-        return None
-    n, unit = int(m.group(1)), m.group(2).lower()
-    return float(n * {"s": 1, "m": 60, "h": 3600}[unit])
-
-
-def is_own_message(sender):
-    s = (sender or "").strip().lower()
-    return s in {x.lower() for x in OWN_SENDERS}
-
-
-def is_authorized(sender):
-    return (sender or "").strip() in AUTHORIZED_SENDERS
-
-
-def extract_loop_commands(messages):
-    """messages: list of {"sender": str, "text": str, "ts": float}.
-    Returns [(sender, verb, args, msg)] for !loop lines from any sender
-    (authorization is applied by the caller so corrections can name names)."""
-    cmds = []
-    for msg in messages:
-        text = msg.get("text") or ""
-        for line in text.splitlines():
-            m = LOOP_CMD_RE.match(line)
-            if m:
-                cmds.append((msg.get("sender", "?"),
-                            m.group(1).lower(),
-                            m.group(2).strip(),
-                            msg))
-    return cmds
-
-
-def prune_expired(state, now=None):
-    """Drop expired ignores / quiet. Returns (notes, changed)."""
-    now = now if now is not None else time.time()
-    notes, changed = [], False
-    ignores = state.setdefault("ignores", {})
-    for agent in list(ignores):
-        if ignores[agent] <= now:
-            del ignores[agent]
-            notes.append("resuming %s (ignore expired)" % agent)
-            changed = True
-    if state.get("quiet_until", 0) and state["quiet_until"] <= now:
-        state["quiet_until"] = 0
-        notes.append("quiet period ended, prompts resumed")
-        changed = True
-    return notes, changed
-
-
-def is_ignored(state, agent, now=None):
-    now = now if now is not None else time.time()
-    return state.get("ignores", {}).get(agent, 0) > now
-
-
-def is_quiet(state, now=None):
-    now = now if now is not None else time.time()
-    return (state.get("quiet_until", 0) or 0) > now
-
-
-def apply_command(sender, verb, args, state, now=None):
-    """Apply one !loop command. Returns (ack_text, changed)."""
-    now = now if now is not None else time.time()
-    changed = False
-
-    if verb == "ignore":
-        parts = args.split()
-        if len(parts) != 2 or parts[0] not in VALID_AGENTS:
-            return ("usage: !loop ignore <agent> <dur>  (agent: %s, dur like 30m/1h)"
-                    % "/".join(sorted(VALID_AGENTS)), False)
-        dur = parse_duration(parts[1])
-        if dur is None or dur <= 0:
-            return ("bad duration %r, try 30m or 1h" % parts[1], False)
-        state.setdefault("ignores", {})[parts[0]] = now + dur
-        return ("ignoring %s for %s (until %s)"
-                % (parts[0], parts[1],
-                   time.strftime("%H:%M UTC", time.gmtime(now + dur))), True)
-
-    if verb == "unignore":
-        agent = args.strip()
-        if agent not in VALID_AGENTS:
-            return ("usage: !loop unignore <agent>", False)
-        if state.get("ignores", {}).pop(agent, None) is not None:
-            changed = True
-            return ("resuming %s now" % agent, True)
-        return ("%s was not ignored" % agent, False)
-
-    if verb == "quiet":
-        dur = parse_duration(args)
-        if dur is None or dur <= 0:
-            return ("usage: !loop quiet <dur>  (dur like 30m/1h)", False)
-        state["quiet_until"] = now + dur
-        changed = True
-        return ("quiet for %s: reads continue, digests land here, "
-                "no per-agent escalation" % args.strip(), True)
-
-    if verb == "unquiet":
-        if state.get("quiet_until"):
-            state["quiet_until"] = 0
-            changed = True
-            return ("prompts resumed", True)
-        return ("was not quiet", False)
-
-    if verb == "check":
-        # The tick loop honors this by running the agent reads immediately
-        # rather than waiting for the next timer fire. Caller sets the flag.
-        return ("CHECK_REQUESTED", True)
-
-    if verb == "status":
-        return ("STATUS_REQUESTED", False)
-
-    return ("unknown command %r, try: ignore, unignore, quiet, unquiet, "
-            "check, status" % verb, False)
-
-
-def format_thinking(tick, summary, state, now=None):
-    """summary: dict with keys seen{agent:(new,q,urgent)}, escalated[JOB ids],
-    skipped_info[int], errors[int], closure_rate[float|None]."""
-    now = now if now is not None else time.time()
-    ts = time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime(now))
-    seen_bits = []
-    for agent in ("646", "pip", "muse", "opm"):
-        new, q, u = summary.get("seen", {}).get(agent, (0, 0, 0))
-        bit = "%s:%dnew" % (agent, new)
-        if q:
-            bit += "(%d?)" % q
-        if u:
-            bit += "(%d!)" % u
-        seen_bits.append(bit)
-    esc = summary.get("escalated", [])
-    lines = [
-        "[BRAIN %s tick=%d]" % (ts, tick),
-        "seen: " + ", ".join(seen_bits),
-        "decided: escalated %d%s | skipped %d info | ignored: %s" % (
-            len(esc),
-            (" (%s)" % ", ".join(esc[:3])) if esc else "",
-            summary.get("skipped_info", 0),
-            ", ".join(sorted(state.get("ignores", {}))) or "none"),
-        "errors: %d | quiet: %s | closure: %s" % (
-            summary.get("errors", 0),
-            "yes" if is_quiet(state, now) else "no",
-            ("%.2f" % summary["closure_rate"])
-            if summary.get("closure_rate") is not None else "n/a"),
-    ]
-    return "\n".join(lines)
-
-
-def format_status(tick, cfg, state, watermarks, health, now=None):
-    now = now if now is not None else time.time()
-    ts = time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime(now))
-    lines = ["[BRAIN %s tick=%d] status" % (ts, tick)]
-    lines.append("agents: " + ", ".join(
-        "%s=%s" % (a, "on" if cfg.get(a) else "off")
-        for a in ("muse", "pip", "646", "opm")))
-    ign = state.get("ignores", {})
-    lines.append("ignores: " + (", ".join(
-        "%s until %s" % (a, time.strftime("%H:%M UTC", time.gmtime(e)))
-        for a, e in sorted(ign.items())) or "none"))
-    q = state.get("quiet_until", 0)
-    lines.append("quiet: " + ("until %s" % time.strftime("%H:%M UTC", time.gmtime(q))
-                              if q and q > now else "no"))
-    ages = []
-    for a in ("muse", "pip", "646", "opm"):
-        wm = (watermarks or {}).get(a, 0)
-        ages.append("%s:%dm" % (a, int((now - wm) / 60)) if wm else "%s:never" % a)
-    lines.append("watermark age: " + ", ".join(ages))
-    if health:
-        lines.append("digest health: delivered=%d acked=%d closed=%d stale=%d "
-                     "closure=%.2f" % (
-                         health.get("delivered", 0), health.get("acked", 0),
-                         health.get("closed", 0), health.get("stale", 0),
-                         health.get("closure_rate", 0.0)))
-    return "\n".join(lines)
-
-
-# ---------------------------------------------------------------------------
-# I/O adapters — thin shells over the bl tools. Verify paths on bl.
-# ---------------------------------------------------------------------------
-
-class BrainIO:
-    """Send/receive against the brain sidechat.
-
-    send: dm.py, WITHOUT --expect-reply (thinking is never actionable).
-    read: pluggable read_fn(messages-since-ts); defaults to None and must be
-          wired to the same chat-read primitive the response-harvester uses.
-    """
-
-    def __init__(self, bin_dir=DEFAULT_BIN_DIR, read_fn=None,
-                 sidechat_map_path=None):
-        self.bin_dir = bin_dir
-        self.read_fn = read_fn
-        self.sidechat_map_path = (sidechat_map_path or
-                                  os.path.join(bin_dir, "..",
-                                               "job-sidechats.json"))
-        self._posts_this_tick = 0
-
-    # -- name resolution: never hardcode the thread UUID -------------------
-    def brain_thread_uuid(self):
-        """Resolve 'main-loop brain' -> UUID via job-sidechats.json."""
-        try:
-            with open(os.path.normpath(self.sidechat_map_path)) as f:
-                data = json.load(f)
-        except (OSError, ValueError):
-            return None
-        # schema: {"sidechats": {"main-loop brain": {"uuid": ...}}} or flat
-        node = data.get("sidechats", data).get(BRAIN_SIDECHAT_NAME)
-        if isinstance(node, dict):
-            return node.get("uuid") or node.get("thread_uuid")
-        return node if isinstance(node, str) else None
-
-    # -- write --------------------------------------------------------------
-    def reset_tick_budget(self):
-        self._posts_this_tick = 0
-
-    def post(self, text):
-        """Post thinking/acks to the brain. Returns True on VERIFIED send."""
-        if self._posts_this_tick >= MAX_BRAIN_POSTS_PER_TICK:
-            return False
-        if JOB_MARKER_RE.search(text):
-            raise ValueError("refusing to post [JOB ...] to the brain")
-        dm = os.path.join(self.bin_dir, "dm.py")
-        # Same path the loop uses for opm digests; NO --expect-reply:
-        # thinking is never actionable and must not create followups.
-        cmd = [dm, "send", "--agent", BRAIN_AGENT,
-               "--to", BRAIN_AGENT, "--target", BRAIN_SIDECHAT_NAME,
-               "--message", text]
-        try:
-            p = subprocess.run(cmd, capture_output=True, text=True,
-                               timeout=120)
-        except (OSError, subprocess.TimeoutExpired):
-            return False
-        ok = "VERIFIED" in (p.stdout or "")
-        if ok:
-            self._posts_this_tick += 1
-        return ok
-
-    # -- read ---------------------------------------------------------------
-    def read_new(self, since_ts):
-        """Return [{"sender","text","ts"}] newer than since_ts. Skips own."""
-        if self.read_fn is None:
-            return []
-        try:
-            msgs = self.read_fn(BRAIN_AGENT, BRAIN_SIDECHAT_NAME, since_ts) or []
-        except Exception:
-            return []
-        return [m for m in msgs
-                if (m.get("ts", 0) or 0) > since_ts
-                and not is_own_message(m.get("sender"))]
-
-
-# ---------------------------------------------------------------------------
-# Workspace — one object per tick
-# ---------------------------------------------------------------------------
-
-class BrainWorkspace:
-    """Owns the brain side of a main-loop tick.
-
-    Usage in self_main_loop.py tick():
-        brain = BrainWorkspace(watermark_path, read_fn=<harvester reader>)
-        cmds_outcome = brain.intake()      # read, parse, apply, ack
-        ... existing per-agent reads, skipping brain.ignored(agent) ...
-        brain.post_thinking(summary)       # the loop's reasoning, visible
-        brain.save()
-    """
-
-    def __init__(self, watermark_path, read_fn=None, bin_dir=DEFAULT_BIN_DIR):
-        self.watermark_path = watermark_path
-        self.io = BrainIO(bin_dir=bin_dir, read_fn=read_fn)
-        self._data = self._load()
-        self.state = self._data.setdefault("brain", {})
-        self.io.reset_tick_budget()
-        self.check_requested = False
-        self.status_requested = False
-
-    # -- persistence ---------------------------------------------------------
-    def _load(self):
-        try:
-            with open(self.watermark_path) as f:
-                return json.load(f)
-        except (OSError, ValueError):
-            return {}
-
-    def save(self):
-        tmp = self.watermark_path + ".tmp"
-        with open(tmp, "w") as f:
-            json.dump(self._data, f, indent=2)
-        os.replace(tmp, self.watermark_path)
-
-    # -- tick intake: read -> parse -> apply -> ack --------------------------
-    def intake(self):
-        """Process new brain messages. Returns dict of what happened."""
-        now = time.time()
-        outcome = {"commands": 0, "acks": 0, "expired_notes": 0,
-                   "ignored_senders": 0}
-        since = float(self.state.get("brain_watermark", 0))
-        msgs = self.io.read_new(since)
-        newest = since
-        for m in msgs:
-            newest = max(newest, float(m.get("ts", 0) or 0))
-        if msgs:
-            self.state["brain_watermark"] = newest
-
-        notes, _ = prune_expired(self.state, now)
-        for n in notes:
-            if self.io.post("[BRAIN %s] %s" % (
-                    time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime(now)), n)):
-                outcome["expired_notes"] += 1
-
-        for sender, verb, args, _msg in extract_loop_commands(msgs):
-            if not is_authorized(sender):
-                outcome["ignored_senders"] += 1
-                continue
-            outcome["commands"] += 1
-            ack, _changed = apply_command(sender, verb, args, self.state, now)
-            if ack == "CHECK_REQUESTED":
-                self.check_requested = True
-                ack = "out-of-cycle check armed for this tick"
-            elif ack == "STATUS_REQUESTED":
-                self.status_requested = True
-                continue  # status posts at end of tick with full context
-            if self.io.post("[BRAIN %s] @%s %s" % (
-                    time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime(now)),
-                    sender, ack)):
-                outcome["acks"] += 1
-        return outcome
-
-    # -- state queries for the tick loop --------------------------------------
-    def ignored(self, agent):
-        return is_ignored(self.state, agent)
-
-    def quiet(self):
-        return is_quiet(self.state)
-
-    # -- end of tick -----------------------------------------------------------
-    def post_thinking(self, summary, cfg=None, watermarks=None, health=None):
-        tick = int(self.state.get("tick", 0)) + 1
-        self.state["tick"] = tick
-        ok = self.io.post(format_thinking(tick, summary, self.state))
-        if self.status_requested and cfg is not None:
-            self.io.post(format_status(tick, cfg, self.state,
-                                       watermarks, health))
-            self.status_requested = False
-        return ok

bin/bridge.py

@@ -1,89 +0,0 @@
-#!/usr/bin/env python3
-"""
-Bridge CLI: Front-Door ↔ muse.ai metadata bridge
-
-Manages mappings between front-door channels and muse.ai side chats.
-Storage: ~/Projects/NetVM/bridge/mappings.json (operator-managed via SSH)
-
-Usage:
-  bridge.py list --channel #ops
-  bridge.py add --agent 646 --chat <id> --channel #ops --purpose "incident-123"
-  bridge.py remove <bridge_id>
-"""
-import json
-import argparse
-from pathlib import Path
-from datetime import datetime, timezone
-
-BRIDGE_DIR = Path.home() / "Projects" / "NetVM" / "bridge"
-MAPPINGS_FILE = BRIDGE_DIR / "mappings.json"
-
-def load_mappings():
-    if not MAPPINGS_FILE.exists():
-        return []
-    with open(MAPPINGS_FILE) as f:
-        return json.load(f)
-
-def save_mappings(mappings):
-    BRIDGE_DIR.mkdir(parents=True, exist_ok=True)
-    with open(MAPPINGS_FILE, 'w') as f:
-        json.dump(mappings, f, indent=2)
-
-def cmd_list(args):
-    mappings = load_mappings()
-    if args.channel:
-        mappings = [m for m in mappings if m['frontdoor_channel'] == args.channel]
-    if not mappings:
-        print("No mappings found.")
-        return
-    for m in mappings:
-        print(f"{m['bridge_id']}: {m['agent']}/{m['muse_side_chat_id'][:8]} -> {m['frontdoor_channel']} ({m['purpose']}) [{m['status']}]")
-
-def cmd_add(args):
-    mappings = load_mappings()
-    bridge_id = f"br-{datetime.now(timezone.utc).strftime('%Y%m%d%H%M%S')}"
-    mapping = {
-        "bridge_id": bridge_id,
-        "frontdoor_channel": args.channel,
-        "muse_side_chat_id": args.chat,
-        "agent": args.agent,
-        "linked_at": datetime.now(timezone.utc).isoformat(),
-        "linked_by": args.by or "operator",
-        "purpose": args.purpose,
-        "status": "active"
-    }
-    mappings.append(mapping)
-    save_mappings(mappings)
-    print(f"Added {bridge_id}")
-
-def cmd_remove(args):
-    mappings = load_mappings()
-    mappings = [m for m in mappings if m['bridge_id'] != args.bridge_id]
-    save_mappings(mappings)
-    print(f"Removed {args.bridge_id}")
-
-def main():
-    p = argparse.ArgumentParser(description="Bridge: front-door to muse.ai")
-    sub = p.add_subparsers(dest='cmd', required=True)
-    
-    p_list = sub.add_parser('list', help='List mappings')
-    p_list.add_argument('--channel', help='Filter by front-door channel')
-    p_list.set_defaults(func=cmd_list)
-    
-    p_add = sub.add_parser('add', help='Add a mapping')
-    p_add.add_argument('--agent', required=True, help='Agent (muse/pip/646)')
-    p_add.add_argument('--chat', required=True, help='muse.ai side chat ID')
-    p_add.add_argument('--channel', required=True, help='Front-door channel (#ops, #lobby, etc.)')
-    p_add.add_argument('--purpose', required=True, help='Purpose of the side chat')
-    p_add.add_argument('--by', help='Who linked (operator/agent)')
-    p_add.set_defaults(func=cmd_add)
-    
-    p_rm = sub.add_parser('remove', help='Remove a mapping')
-    p_rm.add_argument('bridge_id', help='Bridge ID to remove')
-    p_rm.set_defaults(func=cmd_remove)
-    
-    args = p.parse_args()
-    args.func(args)
-
-if __name__ == '__main__':
-    main()

bin/cdp-latency-check.sh

@@ -1,51 +0,0 @@
-#!/bin/bash
-# cdp-latency-check.sh — measure CDP relay latency per node.
-#
-# Probes each node's host-side relay (PEER_IP:CDP_PORT from netvm-names.sh,
-# which pins the registry ports) at /json/version and prints one line per
-# node:
-#   name:latency_ms:code
-# On failure (timeout / connection refused / non-200-ish transport error):
-#   name:FAIL:000
-#
-# Self-contained: only needs netvm-names.sh in the same directory and curl.
-set -u
-
-BIN_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-# shellcheck disable=SC1091
-source "$BIN_DIR/netvm-names.sh"
-
-TIMEOUT_S=10
-
-# Registry-driven node list (was hardcoded 4 nodes; def/dev had no
-# cdp-latency coverage — 2026-10-06).
-watched_nodes() {
-  "$BIN_DIR/netvm-registry.py" 2>/dev/null | cut -d: -f1
-}
-
-# Allow sourcing for tests without running checks.
-if [ "${CDP_LATENCY_CHECK_LIB_ONLY:-}" = "1" ]; then
-  return 0 2>/dev/null || exit 0
-fi
-
-NODES="$(watched_nodes)"
-if [ -z "$NODES" ]; then
-  echo "node registry empty/unreadable" >&2
-  exit 1
-fi
-
-# shellcheck disable=SC2086 (intended word splitting: one node per word)
-for node in $NODES; do
-  netvm_names "$node"
-  url="http://${PEER_IP}:${CDP_PORT}/json/version"
-  probe=$(curl -s -m "$TIMEOUT_S" -o /dev/null -w "%{time_total} %{http_code}" "$url" 2>/dev/null)
-  rc=$?
-  secs=$(printf '%s' "$probe" | awk '{print $1}')
-  code=$(printf '%s' "$probe" | awk '{print $2}')
-  if [ "$rc" -ne 0 ] || [ -z "$code" ] || [ "$code" = "000" ]; then
-    echo "${node}:FAIL:000"
-    continue
-  fi
-  ms=$(awk -v s="$secs" 'BEGIN { printf "%d", (s + 0) * 1000 }')
-  echo "${node}:${ms}:${code}"
-done

bin/cdp-relay-watchdog.sh

@@ -1,145 +0,0 @@
-#!/usr/bin/env bash
-# cdp-relay-watchdog.sh — keep per-node CDP relays alive and correctly routed.
-# Two-stage health check:
-#   1. Host veth IP must be assigned (veth_healthy). Without it the relay is
-#      unreachable from the host no matter how many times we restart it —
-#      observed 2026-10-04 (muse/pip veths existed but had no IPs). FAIL_LOUD
-#      in the log; do NOT auto-fix (veth recreation touches WireGuard/iptables).
-#   2. Relay connectivity (relay_healthy): curl to veth IP:port, not pidfile
-#      (which goes stale and lies — observed 2026-10-04).
-# If a relay is down or misrouted: kill it and restart via the exact
-# netvm-node-up.sh relay invocation inside the node's netns.
-# Runs every 5 min via systemd timer cdp-relay-watchdog.timer.
-# Pattern mirrors chromebox-watchdog.sh (stage-specific logging, rotation).
-set -euo pipefail
-LOCK="/tmp/cdp-relay-watchdog.lock"
-# Tests source this file with CDP_RELAY_WATCHDOG_LIB_ONLY=1: they call
-# helpers without running checks, so no lock is needed.
-if [ "${CDP_RELAY_WATCHDOG_LIB_ONLY:-}" != "1" ]; then
-  exec 9>"$LOCK"
-  if ! flock -n 9; then
-    echo "[$(date -u +%FT%TZ)] another relay watchdog run in progress, skipping" >&2
-    exit 0
-  fi
-fi
-
-NETVM_BIN="/home/super/Projects/NetVM/bin"
-LOG="/home/super/Projects/NetVM/cdp-relay-watchdog.log"
-
-rotate_log() {
-  local f="$1"
-  [ -f "$f" ] || return 0
-  local sz
-  sz=$(stat -c%s "$f" 2>/dev/null || echo 0)
-  if [ "$sz" -gt 10485760 ]; then
-    mv -f "$f" "$f.1"
-    echo "[$(date -u +%FT%TZ)] log rotated" > "$f"
-  fi
-}
-rotate_log "$LOG"
-log() { echo "[$(date -u +%FT%TZ)] $*" | tee -a "$LOG"; }
-
-# node -> "veth_ip:port" via netvm-names.sh (hash-derived, don't hardcode)
-relay_target() {
-  local node="$1" reg_port=""
-  # shellcheck disable=SC1091
-  . "$NETVM_BIN/netvm-names.sh"
-  netvm_names "$node" || return 1
-  # The registry is the source of truth for ports (new nodes propagate
-  # automatically); netvm-names pinning is the fallback.
-  if reg_port=$("$NETVM_BIN/netvm-registry.py" "$node" 2>/dev/null); then
-    [ -n "$reg_port" ] && CDP_PORT="$reg_port"
-  fi
-  echo "$PEER_IP:$CDP_PORT"
-}
-node_port() { echo "${1##*:}"; }
-
-# node -> "VETH GW" via netvm-names.sh
-node_veth() {
-  local node="$1"
-  # shellcheck disable=SC1091
-  . "$NETVM_BIN/netvm-names.sh"
-  netvm_names "$node" || return 1
-  echo "$VETH $GW"
-}
-
-# Is the host-side veth IP assigned? The relay listens on the netns-side peer
-# IP; the host reaches it via the veth interface's GW address. If the GW IP is
-# missing, the relay is unreachable from the host — restarting the relay is
-# pointless and masks the real problem.
-veth_healthy() {
-  local node="$1" veth gw
-  read -r veth gw <<< "$(node_veth "$node")" || return 1
-  ip addr show dev "$veth" 2>/dev/null | grep -q "inet ${gw}/" || return 1
-  return 0
-}
-
-relay_healthy() {
-  local node="$1" target
-  target="$(relay_target "$node")" || return 1
-  curl -s -m 8 "http://$target/json/version" 2>/dev/null | grep -q '"Browser"' || return 1
-  return 0
-}
-
-restart_relay() {
-  local node="$1" target port veth_ip netns
-  target="$(relay_target "$node")"
-  veth_ip="${target%%:*}"
-  port="${target##*:}"
-  netns="warp-$node"
-  # Kill any existing relay for this node's port (correct or not)
-  # Relays are root-owned (started via sudo ip netns exec); the timer runs as
-  # super, so the kill needs sudo too. Without it pkill fails EPERM silently
-  # and the "restart" false-positives via SO_REUSEADDR double-bind.
-  sudo -n pkill -f "netvm-cdp-relay.py .* $port 127.0.0.1 $port" 2>/dev/null || true
-  sleep 2
-  # Launch inside the netns, listening on the veth IP (host-reachable)
-  sudo -n ip netns exec "$netns" setsid nohup python3 \
-    "$NETVM_BIN/netvm-cdp-relay.py" "$veth_ip" "$port" 127.0.0.1 "$port" \
-    >>"$LOG" 2>&1 < /dev/null &
-  sleep 5
-  if relay_healthy "$node"; then
-    log "[$node] relay restarted OK on $target"
-    return 0
-  else
-    log "[$node] relay restart FAILED on $target — needs operator attention"
-    return 1
-  fi
-}
-
-# Registry-driven node list: every active node gets relay supervision
-# (the old hardcoded 4-node list left def/dev unsupervised — 2026-10-06).
-watched_nodes() {
-  "$NETVM_BIN/netvm-registry.py" 2>/dev/null | cut -d: -f1
-}
-
-# Allow sourcing for tests without running checks.
-if [ "${CDP_RELAY_WATCHDOG_LIB_ONLY:-}" = "1" ]; then
-  return 0 2>/dev/null || exit 0
-fi
-
-FAILED=0
-NODES="$(watched_nodes)"
-if [ -z "$NODES" ]; then
-  log "FAIL_LOUD: node registry empty/unreadable, skipping run"
-  exit 1
-fi
-# shellcheck disable=SC2086 (intended word splitting: one node per word)
-for node in $NODES; do
-  # Stage 1: host veth IP. Fail loud, skip relay restart (pointless).
-  if ! veth_healthy "$node"; then
-    read -r veth gw <<< "$(node_veth "$node")"
-    log "[$node] FAIL_LOUD: host veth $veth missing IP $gw — relay unreachable, needs netvm-node-up.sh $node (manual)"
-    FAILED=1
-    continue
-  fi
-  # Stage 2: relay connectivity.
-  if relay_healthy "$node"; then
-    continue
-  fi
-  target="$(relay_target "$node")"
-  log "[$node] relay unhealthy on $target, restarting"
-  restart_relay "$node" || FAILED=1
-done
-
-exit $FAILED

bin/cdp_queue.py

@@ -1,239 +0,0 @@
-#!/usr/bin/env python3
-"""
-Per-browser CDP operation queue for the NetVM fleet.
-
-Problem: nothing coordinates browser operations. DM sends (dm.py), tab
-operations, agent reads (muse-chat-api.py), and watchdog restarts all hit
-the same Chromium with zero scheduling. Functional tests pass under load,
-but there is no backpressure — under real fleet concurrency this will
-overwhelm the machine.
-
-Solution: per-node FIFO queue with priority levels and a cap on concurrent
-CDP operations per browser.
-
-Cross-process design: dm.py shells out to muse-chat-api.py via subprocess,
-so in-process locks (threading.Semaphore) alone cannot coordinate. This
-module uses flock'd slot files + ticket files in /tmp, which work across
-processes AND threads.
-
-Usage:
-    from cdp_queue import cdp_slot, PRIORITY_HIGH
-
-    with cdp_slot("opm", priority=PRIORITY_HIGH):
-        ... do CDP work ...
-
-Explicit acquire/release:
-    from cdp_queue import acquire, release, QueueTimeout
-    token = acquire("opm", priority=PRIORITY_HIGH, timeout=60)
-    try:
-        ...
-    finally:
-        release(token)
-
-Priority levels (lower number = higher priority):
-    PRIORITY_HIGH   = 0  # DM sends (user-facing)
-    PRIORITY_NORMAL = 1  # tab opens, reads
-    PRIORITY_LOW    = 2  # background scans
-
-Fairness: tickets are ordered by (priority, arrival time). A waiter only
-proceeds when its ticket is first in line AND a slot is free. No starvation:
-a low-priority ticket eventually becomes the oldest and gets served.
-"""
-
-import fcntl
-import logging
-import os
-import time
-import uuid
-
-log = logging.getLogger("cdp_queue")
-
-# ---- Tunables ----
-PRIORITY_HIGH = 0
-PRIORITY_NORMAL = 1
-PRIORITY_LOW = 2
-
-MAX_CONCURRENT = 2       # max simultaneous CDP ops per browser
-ACQUIRE_TIMEOUT = 60.0   # fail loud instead of hanging forever
-WARN_AFTER = 10.0        # log a warning when a waiter waits this long
-POLL_INTERVAL = 0.05     # ticket/slot poll cadence
-
-QUEUE_DIR = "/tmp/cdp-queue"
-VALID_NODES = ("muse", "pip", "646", "opm", "dev", "def")
-
-
-class QueueTimeout(Exception):
-    """Raised when a slot cannot be acquired within the timeout."""
-    pass
-
-
-def _node_dir(node):
-    return os.path.join(QUEUE_DIR, node)
-
-
-def _tickets_dir(node):
-    return os.path.join(_node_dir(node), "tickets")
-
-
-def _slot_path(node, i):
-    return os.path.join(_node_dir(node), "slot-%d.lock" % i)
-
-
-def _ensure_dirs(node):
-    os.makedirs(_tickets_dir(node), exist_ok=True)
-    # Pre-create slot files so flock targets always exist
-    for i in range(MAX_CONCURRENT):
-        p = _slot_path(node, i)
-        if not os.path.exists(p):
-            open(p, "a").close()
-
-
-def _read_tickets(node):
-    """Return sorted list of (priority, timestamp, ticket_name), purging stale tickets."""
-    tdir = _tickets_dir(node)
-    out = []
-    now = time.time()
-    try:
-        for name in os.listdir(tdir):
-            if not name.endswith(".ticket"):
-                continue
-            try:
-                # ticket name: "<prio>-<timestamp>-<uuid>.ticket"
-                parts = name[:-7].split("-")
-                prio_s, ts_s = parts[0], parts[1]
-                ts = float(ts_s)
-                # Purge stale ticket if process crashed or timed out ungracefully
-                if now - ts > (ACQUIRE_TIMEOUT * 2):
-                    try:
-                        os.unlink(os.path.join(tdir, name))
-                    except Exception:
-                        pass
-                    continue
-                out.append((int(prio_s), ts, name))
-            except (ValueError, IndexError):
-                continue
-    except FileNotFoundError:
-        pass
-    out.sort()
-    return out
-
-
-class _Slot:
-    """A held queue slot. Release via .release() or context manager."""
-
-    def __init__(self, node, ticket_name, fh, waited):
-        self.node = node
-        self.ticket_name = ticket_name
-        self.fh = fh
-        self.waited = waited
-        self._released = False
-
-    def release(self):
-        if self._released:
-            return
-        self._released = True
-        try:
-            fcntl.flock(self.fh, fcntl.LOCK_UN)
-            self.fh.close()
-        except Exception:
-            pass
-        # Remove our ticket (best effort — a stale ticket is harmless;
-        # the next waiter re-reads the directory each poll)
-        try:
-            os.unlink(os.path.join(_tickets_dir(self.node), self.ticket_name))
-        except Exception:
-            pass
-        log.debug("cdp_queue: released slot for node=%s (waited %.1fs)",
-                  self.node, self.waited)
-
-    def __enter__(self):
-        return self
-
-    def __exit__(self, *exc):
-        self.release()
-
-
-def acquire(node, priority=PRIORITY_NORMAL, timeout=ACQUIRE_TIMEOUT):
-    """
-    Block until a CDP slot is free for `node`, then return a _Slot.
-    Raises QueueTimeout after `timeout` seconds. Raises ValueError for
-    unknown nodes.
-    """
-    if node not in VALID_NODES:
-        raise ValueError("unknown node: %r (valid: %s)" % (node, VALID_NODES))
-    if priority not in (PRIORITY_HIGH, PRIORITY_NORMAL, PRIORITY_LOW):
-        raise ValueError("invalid priority: %r" % (priority,))
-
-    _ensure_dirs(node)
-    tdir = _tickets_dir(node)
-
-    # Our ticket: "<prio>-<timestamp>-<uuid>.ticket", sorted by (prio, ts)
-    ticket = "%d-%f-%s.ticket" % (priority, time.time(), uuid.uuid4().hex[:8])
-    open(os.path.join(tdir, ticket), "w").close()
-
-    start = time.time()
-    warned = False
-    try:
-        while True:
-            elapsed = time.time() - start
-            if elapsed >= timeout:
-                raise QueueTimeout(
-                    "node=%s: no CDP slot free after %.0fs (priority=%d)" %
-                    (node, timeout, priority))
-            if elapsed >= WARN_AFTER and not warned:
-                warned = True
-                depth = len(_read_tickets(node))
-                log.warning("cdp_queue: node=%s waiting %.0fs for slot "
-                            "(queue depth %d, priority %d)",
-                            node, elapsed, depth, priority)
-
-            tickets = _read_tickets(node)
-            # Am I among the first MAX_CONCURRENT in line?
-            # (priority, then arrival time). The first N tickets are all
-            # eligible to grab slots; they distribute via non-blocking flock.
-            my_pos = next((i for i, (_, _, name) in enumerate(tickets)
-                           if name == ticket), None)
-            if my_pos is not None and my_pos < MAX_CONCURRENT:
-                # Try each slot file non-blocking
-                for i in range(MAX_CONCURRENT):
-                    fh = open(_slot_path(node, i), "w")
-                    try:
-                        fcntl.flock(fh, fcntl.LOCK_EX | fcntl.LOCK_NB)
-                    except (BlockingIOError, OSError):
-                        fh.close()
-                        continue
-                    # Got it
-                    waited = time.time() - start
-                    if waited > 1.0:
-                        log.debug("cdp_queue: node=%s acquired slot after "
-                                  "%.1fs (priority %d)", node, waited, priority)
-                    return _Slot(node, ticket, fh, waited)
-            time.sleep(POLL_INTERVAL)
-    except BaseException:
-        # On timeout or interrupt, remove our ticket so we don't block others
-        try:
-            os.unlink(os.path.join(tdir, ticket))
-        except Exception:
-            pass
-        raise
-
-
-def release(slot):
-    """Release a slot returned by acquire()."""
-    slot.release()
-
-
-def cdp_slot(node, priority=PRIORITY_NORMAL, timeout=ACQUIRE_TIMEOUT):
-    """
-    Context manager. Usage:
-        with cdp_slot("opm", priority=PRIORITY_HIGH):
-            ... CDP work ...
-    """
-    return acquire(node, priority=priority, timeout=timeout)
-
-
-def queue_depth(node):
-    """Current number of waiters for a node (for monitoring)."""
-    if node not in VALID_NODES:
-        raise ValueError("unknown node: %r" % node)
-    return len(_read_tickets(node))

bin/chat-rate.py

@@ -1,129 +0,0 @@
-#!/usr/bin/env python3
-"""Chat rate metric: messages per minute for main chat and each side chat.
-
-Polls muse-chat-api.py for message counts, calculates delta vs previous poll,
-logs rates to a time-series file.
-
-Usage: chat-rate.py --account <name> --cdp-port <port> [--interval 60]
-"""
-import json, subprocess, sys, time, os, argparse
-from datetime import datetime, timezone
-
-API = os.path.expanduser("~/Projects/NetVM/bin/muse-chat-api.py")
-STATE_FILE = os.path.expanduser("~/Projects/NetVM/logs/chat-rate-state.json")
-LOG_FILE = os.path.expanduser("~/Projects/NetVM/logs/chat-rate.log")
-
-def run_api(account, *args):
-    """Run muse-chat-api.py and return stdout."""
-    cmd = ["python3", API, "--account", account] + list(args)
-    # Note: cdp-port is baked into the account config, not passed here
-    result = subprocess.run(cmd, capture_output=True, text=True, timeout=60)
-    return result.stdout
-
-def count_messages(text):
-    """Count messages in API output. Messages are separated by '---'."""
-    # The API outputs messages separated by ---\n
-    parts = [p.strip() for p in text.split("---") if p.strip()]
-    # Filter out non-message lines (headers, etc.)
-    # Messages typically have substantial content
-    return len([p for p in parts if len(p) > 10])
-
-def get_side_chats(account):
-    """List side chat names."""
-    out = run_api(account, "sidechat", "list")
-    # Parse: names and timestamps separated by blank lines
-    # Format: "Side chats\n\n<name>\n\n<timestamp>\n\n<name>\n\n<timestamp>..."
-    lines = [l.strip() for l in out.split("\n") if l.strip()]
-    chats = []
-    # Skip header "Side chats", then pair up (name, timestamp)
-    lines = [l for l in lines if l != "Side chats"]
-    # Lines alternate: name, timestamp, name, timestamp...
-    for i in range(0, len(lines), 2):
-        if i < len(lines):
-            name = lines[i]
-            # Verify next is a timestamp (ends with m/h/d)
-            if i + 1 < len(lines) and lines[i+1][-1] in "mhd":
-                chats.append(name)
-    return chats
-
-def get_chat_count(account, chat_name=None):
-    """Get message count for main or a side chat."""
-    if chat_name:
-        # Switch to side chat, get messages, switch back
-        run_api(account, "sidechat", "use", chat_name)
-        out = run_api(account, "messages")
-        run_api(account, "sidechat", "main")  # switch back
-    else:
-        out = run_api(account, "messages")
-    return count_messages(out)
-
-def main():
-    p = argparse.ArgumentParser()
-    p.add_argument("--account", required=True)
-    p.add_argument("--interval", type=int, default=60, help="poll interval seconds")
-    p.add_argument("--once", action="store_true", help="single poll, no loop")
-    args = p.parse_args()
-
-    # Load previous state
-    prev = {}
-    if os.path.exists(STATE_FILE):
-        with open(STATE_FILE) as f:
-            prev = json.load(f)
-
-    def poll():
-        now = datetime.now(timezone.utc).isoformat()
-        counts = {}
-        
-        # Main chat
-        try:
-            counts["main"] = get_chat_count(args.account)
-        except Exception as e:
-            print(f"main: error {e}", file=sys.stderr)
-        
-        # Side chats
-        try:
-            sc_list = get_side_chats(args.account)
-            print(f"DEBUG: found {len(sc_list)} side chats", file=sys.stderr)
-            for sc in sc_list:
-                try:
-                    counts[f"side:{sc}"] = get_chat_count(args.account, sc)
-                except Exception as e:
-                    print(f"side:{sc}: error {e}", file=sys.stderr)
-        except Exception as e:
-            print(f"sidechat list: error {e}", file=sys.stderr)
-        
-        # Calculate rates
-        results = []
-        for chat, count in counts.items():
-            rate = 0.0
-            if chat in prev:
-                prev_count, prev_time = prev[chat]
-                dt = (datetime.fromisoformat(now) - datetime.fromisoformat(prev_time)).total_seconds() / 60.0
-                if dt > 0:
-                    rate = (count - prev_count) / dt
-            results.append((now, chat, count, round(rate, 2)))
-            prev[chat] = (count, now)
-        
-        # Log
-        os.makedirs(os.path.dirname(LOG_FILE), exist_ok=True)
-        with open(LOG_FILE, "a") as f:
-            for ts, chat, count, rate in results:
-                f.write(f"{ts} {chat} count={count} rate={rate}/min\n")
-        
-        # Save state
-        with open(STATE_FILE, "w") as f:
-            json.dump(prev, f)
-        
-        # Print
-        for ts, chat, count, rate in results:
-            print(f"{chat}: {count} msgs, {rate}/min")
-
-    if args.once:
-        poll()
-    else:
-        while True:
-            poll()
-            time.sleep(args.interval)
-
-if __name__ == "__main__":
-    main()

bin/chat-state-check.py

@@ -1,281 +0,0 @@
-#!/usr/bin/env python3
-"""chat-state-check.py — one-shot CDP chat-state probe for a single node.
-
-Runs INSIDE the node's netns (CDP listens on 127.0.0.1 there).
-
-Usage: chat-state-check.py <cdp_port> <mode> [name]
-  modes: main | list | sidechat <name>
-
-Prints one JSON object to stdout.
-
-Honest limits (see CHATSTATE_SPEC.md): only the DOM-visible message window
-is captured (React virtualization); author attribution is best-effort and
-may be null; checked_at is the report time, not per-message times.
-"""
-import base64
-import json
-import sys
-import time
-import urllib.request
-
-import websocket
-
-MSG_N = 10
-MSG_WIDTH = 500
-
-
-def connect(port):
-    with urllib.request.urlopen(
-            "http://127.0.0.1:%s/json/list" % port, timeout=5) as r:
-        targets = json.load(r)
-    pages = [t for t in targets if t.get("type") == "page"]
-    if not pages:
-        return None
-    return websocket.create_connection(
-        pages[0]["webSocketDebuggerUrl"], timeout=15)
-
-
-def ev1(ws, expr, await_p=False, reads=30):
-    """Runtime.evaluate that skips CDP event chatter while awaiting ours."""
-    ws.send(json.dumps({
-        "id": 1, "method": "Runtime.evaluate",
-        "params": {"expression": expr, "returnByValue": True,
-                   "awaitPromise": await_p}}))
-    for _ in range(reads):
-        resp = json.loads(ws.recv())
-        if resp.get("id") != 1:
-            continue
-        if "error" in resp:
-            raise RuntimeError("CDP evaluate failed: %s" % resp["error"])
-        res = resp.get("result", {}).get("result", {})
-        return res.get("value")
-    raise RuntimeError("CDP evaluate: no response")
-
-
-def read_current(ws):
-    title = ev1(ws, "document.title") or ""
-    url = ev1(ws, "window.location.href") or ""
-    msgs = ev1(ws, """(() => {
-      const ps = [...document.querySelectorAll('p')].slice(-%d)
-        .map(p => (p.innerText||'').slice(0,%d)).filter(t => t.trim());
-      return ps;
-    })()""" % (MSG_N, MSG_WIDTH)) or []
-    return {"title": title, "url": url,
-            "messages": [{"text": t} for t in msgs]}
-
-
-def _wait_for(ws, expr, timeout_s=12):
-    """Poll a JS truthiness expression until true or timeout."""
-    deadline = time.time() + timeout_s
-    while time.time() < deadline:
-        try:
-            if ev1(ws, expr):
-                return True
-        except Exception:  # noqa: BLE001
-            pass
-        time.sleep(1)
-    return False
-
-
-def ensure_sidebar_open(ws):
-    """The toggle closes an open sidebar — only click when closed.
-
-    Detected structurally via the 'New side chat' button (text matching
-    'Side chats' is unreliable: main-chat messages can contain the phrase).
-    Waits for React to render the panel after opening.
-    """
-    is_open = ev1(ws, """(() => {
-      return !!document.querySelector('button[aria-label="New side chat"]');
-    })()""")
-    if not is_open:
-        ev1(ws, """(() => {
-          const btn = [...document.querySelectorAll('button')].find(b =>
-            (b.textContent||'').includes('Open chat and side chats'));
-          if (btn) btn.click();
-        })()""")
-        _wait_for(ws, """(() => {
-          return !!document.querySelector(
-            'button[aria-label="New side chat"]');
-        })()""")
-    # The row list renders a beat after the panel; wait for it.
-    _wait_for(ws, """(() => {
-      return document.querySelectorAll(
-        'button[aria-label="More thread actions"]').length > 0;
-    })()""", timeout_s=10)
-
-
-_SIDEBAR_JS = """(() => {
-  const anchor = document.querySelector('button[aria-label="New side chat"]');
-  if (!anchor) return 'NOANCHOR';
-  let sec = anchor.parentElement;
-  for (let i = 0; i < 8 && sec; i++) {
-    if (sec.querySelectorAll(
-        'button[aria-label="More thread actions"]').length) break;
-    sec = sec.parentElement;
-  }
-  if (!sec) return 'NOSEC';
-  return JSON.stringify({ok: true});
-})()"""
-
-
-def _sidebar_section(ws):
-    """Return True when the side-chat list section is addressable."""
-    raw = ev1(ws, _SIDEBAR_JS)
-    try:
-        return json.loads(raw or "").get("ok", False)
-    except (ValueError, TypeError):
-        return False
-
-
-_LIST_JS = """(() => {
-  const T = (el) => (el.innerText || '').trim();
-  const TS =
-    /^(\\d+[smhd]|just now|Yesterday|Today|[A-Z][a-z]{2} \\d{1,2}.*)$/;
-  const Y = (el) => el.getBoundingClientRect().top;
-  const h3s = [...document.querySelectorAll('h3')];
-  const head = h3s.find(h => h.textContent.trim() === 'Side chats');
-  if (!head) return '[]';
-  // Scroll the list to top so the section's rows sit above the next
-  // (sticky) header; without this, rows below the fold are missed.
-  let sc = head.parentElement;
-  for (let i = 0; i < 8 && sc; i++) {
-    if (sc.scrollHeight > sc.clientHeight + 10) break;
-    sc = sc.parentElement;
-  }
-  if (sc) sc.scrollTop = 0;
-  const y0 = Y(head);
-  const isStopText = (t) => t === 'Unread updates' || t === 'Chats' ||
-    t === 'Unread chats';
-  let stopY = null;
-  for (const h of h3s) {
-    const y = Y(h);
-    if (y > y0 + 5 && (stopY === null || y < stopY)) stopY = y;
-  }
-  // Non-H3 section headers ("Unread updates", ...) also bound the list.
-  for (const el of document.querySelectorAll('*')) {
-    if (el.children.length !== 0) continue;
-    if (!isStopText((el.textContent || '').trim())) continue;
-    const y = Y(el);
-    if (y > y0 + 5 && (stopY === null || y < stopY)) stopY = y;
-  }
-  const names = [];
-  for (const el of document.querySelectorAll('div')) {
-    const cn = (el.className || '').toString();
-    if (cn.indexOf('nav-row') === -1) continue;
-    const y = Y(el);
-    if (y <= y0 + 5) continue;
-    if (stopY !== null && y >= stopY - 5) continue;
-    const m = T(el).match(/^([^\\n]+)\\n\\n(.+)$/);
-    if (m && TS.test(m[2].trim()) && names.indexOf(m[1].trim()) === -1) {
-      names.push(m[1].trim());
-      if (names.length >= 8) break;
-    }
-  }
-  return JSON.stringify(names);
-})()"""
-
-
-def _list_once(ws):
-    raw = ev1(ws, _LIST_JS)
-    try:
-        return json.loads(raw or "[]")
-    except (ValueError, TypeError):
-        return []
-
-
-def list_sidechats(ws):
-    ensure_sidebar_open(ws)
-    if not _sidebar_section(ws):
-        return []
-    # React renders rows progressively after navigation; poll until the
-    # list stabilizes instead of trusting the first paint.
-    prev = None
-    for _ in range(4):
-        names = _list_once(ws)
-        if names and names == prev:
-            return names
-        prev = names
-        time.sleep(2)
-    return prev or []
-
-
-def open_sidechat(ws, name):
-    """Open the side chat by clicking its row div inside the sidebar section.
-
-    Clicking the row (not a text search over the whole document) avoids
-    hitting message text that happens to contain the chat name.
-    """
-    ensure_sidebar_open(ws)
-    b64 = base64.b64encode(name.encode()).decode()
-    return ev1(ws, """(async () => {
-      const nm = atob('%s');
-      const T = (el) => (el.innerText || '').trim();
-      const target = [...document.querySelectorAll('div')].find(el => {
-        const cn = (el.className || '').toString();
-        if (cn.indexOf('nav-row') === -1) return false;
-        const m = T(el).match(/^([^\\n]+)\\n\\n(.+)$/);
-        return m && m[1].trim() === nm;
-      });
-      if (!target) return 'NOTFOUND';
-      target.click();
-      await new Promise(r => setTimeout(r, 4000));
-      const href = window.location.href;
-      if (href.indexOf('/thread/') === -1) return 'NOTFOUND';
-      return href;
-    })()""" % b64, await_p=True)
-
-
-def navigate_main(ws):
-    ws.send(json.dumps({"id": 2, "method": "Page.navigate",
-                        "params": {"url": "https://muse.ai/"}}))
-    for _ in range(20):
-        resp = json.loads(ws.recv())
-        if resp.get("id") == 2:
-            break
-    time.sleep(6)
-
-
-def main():
-    if len(sys.argv) < 3:
-        print(json.dumps({"ok": False, "error": "usage"}))
-        sys.exit(1)
-    port, mode = sys.argv[1], sys.argv[2]
-    name = sys.argv[3] if len(sys.argv) > 3 else ""
-    try:
-        ws = connect(port)
-    except Exception as e:  # noqa: BLE001
-        print(json.dumps({"ok": False, "error": "cdp_connect_failed",
-                          "detail": str(e)[:120]}))
-        return
-    if ws is None:
-        print(json.dumps({"ok": False, "error": "no_page"}))
-        return
-    try:
-        if mode == "list":
-            print(json.dumps({"ok": True,
-                              "sidechats": list_sidechats(ws)}))
-        elif mode == "sidechat":
-            href = open_sidechat(ws, name)
-            if href == "NOTFOUND":
-                print(json.dumps({"ok": False, "error": "sidechat_not_found",
-                                  "name": name}))
-            else:
-                cur = read_current(ws)
-                cur.update({"ok": True, "context": "sidechat", "name": name})
-                print(json.dumps(cur))
-        else:
-            navigate_main(ws)
-            cur = read_current(ws)
-            cur.update({"ok": True, "context": "main", "name": "Main chat"})
-            print(json.dumps(cur))
-    except Exception as e:  # noqa: BLE001
-        print(json.dumps({"ok": False, "error": "probe_failed",
-                          "detail": str(e)[:200]}))
-    finally:
-        try:
-            ws.close()
-        except Exception:  # noqa: BLE001
-            pass
-
-
-main()

bin/chat-state-report.py

@@ -1,129 +0,0 @@
-#!/usr/bin/env python3
-"""chat-state-report.py — per-node chat-state tap for box.
-
-Probes each NetVM node's browser via CDP (inside its netns): main chat +
-side chats, last messages. Signs and POSTs to the board chat-state ingest,
-mirroring accounts-health.sh: payload is
-<machine>\\n<ts>\\n<facts-json>, namespace "health".
-
-Usage: chat-state-report.py [--no-post]
-
-Cron (on bl, every 5 min):
-  */5 * * * * ~/Projects/NetVM/bin/chat-state-report.py >/dev/null 2>&1
-
-Health key setup: same key as accounts-health (namespace "health",
-registered in /srv/board/health_signers for machine bl).
-"""
-import importlib.util
-import json
-import os
-import subprocess
-import sys
-import tempfile
-import time
-
-NETVM_DIR = os.environ.get("NETVM_DIR",
-                           os.path.expanduser("~/Projects/NetVM"))
-CHECK = os.path.join(NETVM_DIR, "bin", "chat-state-check.py")
-MACHINE = os.environ.get("MUSE_MACHINE", "bl")
-KEY = os.environ.get("CHATSTATE_KEY",
-                     os.path.expanduser("~/.ssh/muse-health"))
-ENDPOINT = os.environ.get(
-    "CHATSTATE_ENDPOINT",
-    "https://board.muse-dev.online/api/box/chat-state/report")
-MAX_SIDECHATS = 5
-FACTS_CAP = 64 * 1024
-POST = "--no-post" not in sys.argv
-
-
-def load_nodes():
-    spec = importlib.util.spec_from_file_location(
-        "netvm_registry",
-        os.path.join(NETVM_DIR, "bin", "netvm-registry.py"))
-    mod = importlib.util.module_from_spec(spec)
-    spec.loader.exec_module(mod)
-    return mod.load()  # node -> {"cdp_port": ...}
-
-
-def run_check(node, port, *args):
-    cmd = ["sudo", "-n", "ip", "netns", "exec", "warp-%s" % node,
-           "python3", CHECK, str(port)] + list(args)
-    try:
-        r = subprocess.run(cmd, capture_output=True, text=True, timeout=150)
-        line = r.stdout.strip().splitlines()[-1]
-        return json.loads(line)
-    except Exception as e:  # noqa: BLE001
-        return {"ok": False, "error": "check_failed",
-                "detail": str(e)[:120]}
-
-
-def main():
-    nodes = load_nodes()
-    chats = []
-    now = int(time.time())
-    for node, rec in sorted(nodes.items()):
-        port = rec.get("cdp_port")
-        if not port:
-            continue
-        base = {"node": node, "checked_at": now}
-        m = run_check(node, port, "main")
-        entry = dict(base)
-        if m.get("ok"):
-            entry.update(m)
-        else:
-            entry.update({"context": "main", "error": m.get("error"),
-                          "detail": m.get("detail", "")[:120]})
-            chats.append(entry)
-            continue
-        chats.append(entry)
-        lst = run_check(node, port, "list")
-        for name in (lst.get("sidechats") or [])[:MAX_SIDECHATS]:
-            s = run_check(node, port, "sidechat", name)
-            sentry = dict(base)
-            if s.get("ok"):
-                sentry.update(s)
-            else:
-                sentry.update({"context": "sidechat", "name": name,
-                               "error": s.get("error"),
-                               "detail": s.get("detail", "")[:120]})
-            chats.append(sentry)
-    facts = {"chats": chats, "checked_at": now}
-    facts_json = json.dumps(facts, separators=(",", ":"))
-    if len(facts_json) > FACTS_CAP:
-        # Shed message bodies, keep metadata.
-        for c in chats:
-            c.pop("messages", None)
-        facts["truncated"] = True
-        facts_json = json.dumps(facts, separators=(",", ":"))
-    if not POST:
-        print(json.dumps(json.loads(facts_json), indent=2))
-        return
-    if not os.path.exists(KEY):
-        print("chat-state-report: %s missing — printing JSON, not posting"
-              % KEY, file=sys.stderr)
-        print(facts_json)
-        return
-    tmp = tempfile.mkdtemp()
-    try:
-        payload = os.path.join(tmp, "payload")
-        with open(payload, "w") as f:
-            f.write("%s\n%d\n%s" % (MACHINE, now, facts_json))
-        # Fresh temp dir: no stale .sig to worry about (cf. AGENTS.md).
-        subprocess.run(["ssh-keygen", "-Y", "sign", "-f", KEY,
-                        "-n", "health", payload],
-                       check=True, capture_output=True)
-        with open(payload + ".sig") as f:
-            sig = f.read()
-        body = {"machine": MACHINE, "ts": now, "facts_json": facts_json,
-                "signature": sig}
-        r = subprocess.run(
-            ["curl", "-s", "-X", "POST", ENDPOINT,
-             "-H", "Content-Type: application/json",
-             "--data", json.dumps(body)],
-            capture_output=True, text=True, timeout=90)
-        print(r.stdout[:300])
-    finally:
-        subprocess.run(["rm", "-rf", tmp])
-
-
-main()

bin/chrome-error-scan.sh

@@ -1,99 +0,0 @@
-#!/bin/bash
-# chrome-error-scan.sh - scan per-profile chrome logs for concerning patterns.
-# Self-contained: scans, compares against watermark, reports only NEW matches.
-#
-# Usage: chrome-error-scan.sh [--json] [--no-advance]
-#   Default output: "profile:new_count" lines for profiles with new matches,
-#                   or "OK: no new errors" if clean.
-#   --json: output JSON {"profile": {"total": N, "new": M}, ...}
-#   --no-advance: report against the watermark WITHOUT advancing it.
-#                 Peek-only read for high-frequency pollers (e.g. the web
-#                 surface via `box-ctl chrome-errors --no-advance`). Runs
-#                 without the flag keep the classic advance-on-read semantics.
-#
-# Watermark: /home/super/Projects/NetVM/chrome-error-watermark.json
-# Patterns: FATAL, crash, segfault, out of memory (case-insensitive)
-
-LOGDIR="/home/super/Projects/NetVM"
-WATERMARK="$LOGDIR/chrome-error-watermark.json"
-
-AS_JSON=0
-NO_ADVANCE=0
-for arg in "$@"; do
-  case "$arg" in
-    --json) AS_JSON=1 ;;
-    --no-advance) NO_ADVANCE=1 ;;
-    *) echo "chrome-error-scan.sh: unknown argument: $arg" >&2; exit 2 ;;
-  esac
-done
-
-# Gather current counts per profile (grep -c prints 0 with exit 1 on no match;
-# the || true masks the exit code while preserving the "0" on stdout)
-get_count() {
-  local f="$LOGDIR/chromebox-$1.log"
-  if [ -f "$f" ]; then
-    grep -ciE "FATAL|crash|segfault|out of memory" "$f" 2>/dev/null || true
-  else
-    echo 0
-  fi
-}
-
-MUSE_C=$(get_count muse)
-PIP_C=$(get_count pip)
-N646_C=$(get_count 646)
-OPM_C=$(get_count opm)
-
-python3 - "$WATERMARK" "$MUSE_C" "$PIP_C" "$N646_C" "$OPM_C" "$AS_JSON" "$NO_ADVANCE" <<'PYEOF'
-import json, sys, os
-
-watermark_path = sys.argv[1]
-current = {
-    "muse": int(sys.argv[2]),
-    "pip": int(sys.argv[3]),
-    "646": int(sys.argv[4]),
-    "opm": int(sys.argv[5]),
-}
-as_json = len(sys.argv) > 6 and sys.argv[6] == "1"
-no_advance = len(sys.argv) > 7 and sys.argv[7] == "1"
-
-# Load watermark (tolerate missing/corrupt file -> treat as all-zero)
-watermark = {}
-if os.path.exists(watermark_path):
-    try:
-        with open(watermark_path) as f:
-            data = json.load(f)
-            watermark = data.get("counts", data) if isinstance(data, dict) else {}
-    except (ValueError, IOError, OSError):
-        watermark = {}
-
-new_counts = {}
-for prof in ("muse", "pip", "646", "opm"):
-    old = watermark.get(prof, 0)
-    try:
-        old = int(old)
-    except (TypeError, ValueError):
-        old = 0
-    delta = current[prof] - old
-    new_counts[prof] = max(delta, 0)
-
-if as_json:
-    out = {p: {"total": current[p], "new": new_counts[p]} for p in current}
-    print(json.dumps(out))
-else:
-    any_new = False
-    for prof in ("muse", "pip", "646", "opm"):
-        if new_counts[prof] > 0:
-            print("%s:%d" % (prof, new_counts[prof]))
-            any_new = True
-    if not any_new:
-        print("OK: no new errors")
-
-# Update watermark atomically (skipped in --no-advance peek mode: the
-# caller gets a read-only view and the CLI's advance-on-read semantics are
-# left untouched).
-if not no_advance:
-    tmp = watermark_path + ".tmp"
-    with open(tmp, "w") as f:
-        json.dump({"counts": current}, f, indent=2)
-    os.replace(tmp, watermark_path)
-PYEOF

bin/chromebox-gateway.py

@@ -1,364 +0,0 @@
-#!/usr/bin/env python3
-"""
-chromebox-gateway.py — HTTPS fallback for chromebox/DM control when SSH is down.
-
-The chromeboxes (headless Chromium per NetVM node) are normally driven over
-SSH via muse-chat-api.py / dm.py, which speak CDP through the per-node
-netns relay. When SSH to bl breaks, this gateway provides a constrained
-HTTPS path to the same high-level operations.
-
-CRITICAL: this does NOT expose raw CDP. Raw CDP (Runtime.evaluate,
-Page.navigate, etc.) is arbitrary code execution inside the browser with
-the agent's live session. This gateway exposes ONLY the allowlisted
-high-level operations below, each mapped to an existing audited script.
-
-Usage:
-    python3 chromebox-gateway.py --port 8444
-
-Auth: Bearer token, reusing the shared exec per-agent token files
-(~/.exec-tokens/<agent>). The master token (~/.exec-server-token) also
-works. Identity = token filename; tokens are never logged.
-
-Endpoints:
-    GET  /health          {"status":"ok"} — no auth (load-balancer friendly)
-    POST /api/v1/op       {"op": "<name>", "params": {...}} — bearer auth
-
-Audit: every call appended to ~/.chromebox-gateway-audit.jsonl (0600).
-"""
-import argparse
-import hmac
-import json
-import os
-import re
-import ssl
-import subprocess
-import sys
-import time
-from http.server import HTTPServer, BaseHTTPRequestHandler
-from urllib.parse import urlparse
-
-# ---------------------------------------------------------------- config
-
-BIN_DIR = os.path.expanduser("~/Projects/NetVM/bin")
-TOKEN_FILE = "/home/super/.exec-server-token"   # master token (shared exec token file)
-TOKEN_DIR = "/home/super/.exec-tokens"          # per-agent tokens (shared exec token files)
-AUDIT_FILE = "/home/super/.chromebox-gateway-audit.jsonl"
-CERT_FILE = "/home/super/.chromebox-gateway-cert.pem"
-KEY_FILE = "/home/super/.chromebox-gateway-key.pem"
-
-NODES = ("muse", "pip", "646", "opm")
-MAX_MSG = 2000          # gateway-level cap; downstream scripts enforce their own
-BACKEND_TIMEOUT = 120   # seconds per backend call
-
-# Rate limit: token bucket per identity — 20 req/min sustained, burst 5.
-RATE_PER_SEC = 20.0 / 60.0
-RATE_BURST = 5
-
-# ---------------------------------------------------------------- ops
-# Each op maps to an argv builder for an existing script. No shell=True,
-# ever. Params are validated before building argv.
-
-def _node(params):
-    node = params.get("account") or params.get("agent")
-    if node not in NODES:
-        raise ValueError("account/agent must be one of %s" % (",".join(NODES)))
-    return node
-
-def _msg(params):
-    m = params.get("message", "")
-    if not isinstance(m, str) or not m.strip():
-        raise ValueError("message must be a non-empty string")
-    if len(m) > MAX_MSG:
-        raise ValueError("message exceeds %d chars" % MAX_MSG)
-    return m
-
-def _target(params):
-    t = params.get("target", "main")
-    if not isinstance(t, str) or not t or len(t) > 128:
-        raise ValueError("target must be a short string")
-    if not re.fullmatch(r"[A-Za-z0-9_./:-]+", t):
-        raise ValueError("target has invalid characters")
-    if ".." in t:
-        raise ValueError("target must not contain '..'")
-    return t
-
-def _n(params, default=5, cap=50):
-    n = params.get("n", default)
-    try:
-        n = int(n)
-    except (TypeError, ValueError):
-        raise ValueError("n must be an integer")
-    if not 1 <= n <= cap:
-        raise ValueError("n must be 1..%d" % cap)
-    return n
-
-def _tags(params):
-    tags = params.get("tags", [])
-    if not isinstance(tags, list):
-        raise ValueError("tags must be a list")
-    out = []
-    for t in tags:
-        if not isinstance(t, str) or len(t) > 128:
-            raise ValueError("bad tag")
-        if not re.fullmatch(r"[A-Za-z0-9_:=\-./]+", t):
-            raise ValueError("tag has invalid characters: %r" % t[:40])
-        out.append(t)
-    if len(out) > 10:
-        raise ValueError("too many tags (max 10)")
-    return out
-
-CHAT = os.path.join(BIN_DIR, "muse-chat-api.py")
-DM = os.path.join(BIN_DIR, "dm.py")
-
-def op_chat_send(p):
-    return [CHAT, "--account", _node(p), "send", _msg(p)]
-
-def op_chat_messages(p):
-    return [CHAT, "--account", _node(p), "messages", "--n", str(_n(p))]
-
-def op_chat_sidechats(p):
-    return [CHAT, "--account", _node(p), "sidechat", "list"]
-
-def op_chat_sidechat_create(p):
-    argv = [CHAT, "--account", _node(p), "sidechat", "create"]
-    name = p.get("name")
-    if name:
-        if not isinstance(name, str) or len(name) > 80 or not re.fullmatch(r"[A-Za-z0-9 _-]+", name):
-            raise ValueError("bad sidechat name")
-        argv += ["--name", name]
-    return argv
-
-def op_chat_approvals(p):
-    return [CHAT, "--account", _node(p), "approvals"]
-
-def op_chat_url(p):
-    return [CHAT, "--account", _node(p), "url"]
-
-def op_dm_send(p):
-    node = _node(p)
-    to = p.get("to", node)
-    if to not in NODES:
-        raise ValueError("to must be one of %s" % (",".join(NODES)))
-    argv = [DM, "send", "--agent", node, "--to", to,
-            "--target", _target(p), _msg(p)]
-    for t in _tags(p):
-        argv += ["--tag", t]
-    return argv
-
-def op_dm_read(p):
-    return [DM, "read", "--agent", _node(p),
-            "--target", _target(p), "--n", str(_n(p, default=5, cap=20))]
-
-# The allowlist. Adding an op here is a security decision — review accordingly.
-# Deliberately absent: wait (long-poll), upload (file ingress),
-# sidechat use (raw navigation), anything raw-CDP.
-ALLOWLIST = {
-    "chat.send":            ("write", op_chat_send),
-    "chat.messages":        ("read",  op_chat_messages),
-    "chat.sidechats":       ("read",  op_chat_sidechats),
-    "chat.sidechat_create": ("write", op_chat_sidechat_create),
-    "chat.approvals":       ("read",  op_chat_approvals),
-    "chat.url":             ("read",  op_chat_url),
-    "dm.send":              ("write", op_dm_send),
-    "dm.read":              ("read",  op_dm_read),
-}
-
-# ---------------------------------------------------------------- auth
-
-def _read_token_file(path):
-    try:
-        with open(path) as f:
-            return f.read().strip()
-    except OSError:
-        return ""
-
-def check_token(token):
-    """Return identity label or None. Tokens never leave this function."""
-    if not token:
-        return None
-    if hmac.compare_digest(token, _read_token_file(TOKEN_FILE)):
-        return "master"
-    try:
-        names = os.listdir(TOKEN_DIR)
-    except OSError:
-        return None
-    for name in names:
-        if not re.fullmatch(r"[A-Za-z0-9_-]+", name):
-            continue
-        t = _read_token_file(os.path.join(TOKEN_DIR, name))
-        if t and hmac.compare_digest(token, t):
-            return name
-    return None
-
-# ---------------------------------------------------------------- rate limit
-
-_buckets = {}  # identity -> [tokens, last_ts]
-
-def rate_ok(identity):
-    now = time.monotonic()
-    tokens, last = _buckets.get(identity, (RATE_BURST, now))
-    tokens = min(RATE_BURST, tokens + (now - last) * RATE_PER_SEC)
-    if tokens < 1.0:
-        _buckets[identity] = (tokens, now)
-        return False
-    _buckets[identity] = (tokens - 1.0, now)
-    return True
-
-# ---------------------------------------------------------------- audit
-
-def audit(entry):
-    entry = dict(entry)
-    entry["ts"] = time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime())
-    # Defense in depth: never let a full message body reach the audit file,
-    # even if a caller forgets to summarize first.
-    params = entry.get("params")
-    if isinstance(params, dict) and "message" in params:
-        params = dict(params)
-        params["message_len"] = len(params.pop("message"))
-        entry["params"] = params
-    try:
-        fd = os.open(AUDIT_FILE, os.O_WRONLY | os.O_CREAT | os.O_APPEND, 0o600)
-        with os.fdopen(fd, "a") as f:
-            f.write(json.dumps(entry) + "\n")
-    except OSError as e:
-        print("audit write failed: %s" % e, file=sys.stderr)
-
-# ---------------------------------------------------------------- handler
-
-class Handler(BaseHTTPRequestHandler):
-    server_version = "chromebox-gateway/1.0"
-
-    def log_message(self, fmt, *args):  # quiet; audit log is the record
-        pass
-
-    def _json(self, code, obj):
-        body = json.dumps(obj).encode()
-        self.send_response(code)
-        self.send_header("Content-Type", "application/json")
-        self.send_header("Content-Length", str(len(body)))
-        self.end_headers()
-        self.wfile.write(body)
-
-    def do_GET(self):
-        if urlparse(self.path).path == "/health":
-            self._json(200, {"status": "ok", "ops": sorted(ALLOWLIST)})
-            return
-        self._json(404, {"error": "not_found"})
-
-    def do_POST(self):
-        if urlparse(self.path).path != "/api/v1/op":
-            self._json(404, {"error": "not_found"})
-            return
-
-        auth = self.headers.get("Authorization", "")
-        token = auth[7:] if auth.startswith("Bearer ") else ""
-        identity = check_token(token)
-        if not identity:
-            self._json(401, {"error": "unauthorized"})
-            return
-        if not rate_ok(identity):
-            audit({"identity": identity, "op": None, "result": "rate_limited"})
-            self._json(429, {"error": "rate_limited"})
-            return
-
-        try:
-            length = int(self.headers.get("Content-Length", 0))
-        except ValueError:
-            length = 0
-        if length > 65536:
-            self._json(413, {"error": "body_too_large"})
-            return
-        try:
-            req = json.loads(self.rfile.read(length) or b"{}")
-        except (ValueError, OSError):
-            self._json(400, {"error": "bad_json"})
-            return
-
-        op = req.get("op")
-        params = req.get("params") or {}
-        if not isinstance(params, dict):
-            self._json(400, {"error": "params_must_be_object"})
-            return
-        entry = ALLOWLIST.get(op)
-        if not entry:
-            audit({"identity": identity, "op": op, "result": "unknown_op"})
-            self._json(400, {"error": "unknown_op", "allowed": sorted(ALLOWLIST)})
-            return
-        cls, builder = entry
-
-        t0 = time.monotonic()
-        try:
-            argv = builder(params)
-        except ValueError as e:
-            audit({"identity": identity, "op": op, "class": cls, "result": "bad_params",
-                   "detail": str(e)[:120]})
-            self._json(400, {"error": "bad_params", "detail": str(e)})
-            return
-
-        # Redacted summary for the audit log — never the full message.
-        summary = {k: (v[:80] + "…" if isinstance(v, str) and len(v) > 80 else v)
-                   for k, v in params.items() if k != "message"}
-        if "message" in params:
-            summary["message_len"] = len(params["message"])
-
-        try:
-            proc = subprocess.run(argv, capture_output=True, text=True,
-                                  timeout=BACKEND_TIMEOUT)
-            ok = proc.returncode == 0
-            result = "ok" if ok else "backend_error"
-            self._json(200 if ok else 502, {
-                "ok": ok,
-                "op": op,
-                "returncode": proc.returncode,
-                "stdout": proc.stdout[-8000:],
-                "stderr": proc.stderr[-2000:],
-            })
-        except subprocess.TimeoutExpired:
-            result = "timeout"
-            self._json(504, {"ok": False, "op": op, "error": "backend_timeout"})
-        except OSError as e:
-            result = "exec_failed"
-            self._json(500, {"ok": False, "op": op, "error": "exec_failed"})
-        finally:
-            audit({"identity": identity, "op": op, "class": cls,
-                   "node": params.get("account") or params.get("agent"),
-                   "params": summary, "result": result,
-                   "latency_ms": int((time.monotonic() - t0) * 1000)})
-
-# ---------------------------------------------------------------- main
-
-def ensure_cert():
-    if os.path.exists(CERT_FILE) and os.path.exists(KEY_FILE):
-        return
-    print("generating self-signed cert...", file=sys.stderr)
-    subprocess.run([
-        "openssl", "req", "-x509", "-newkey", "rsa:2048",
-        "-keyout", KEY_FILE, "-out", CERT_FILE,
-        "-days", "825", "-nodes", "-subj", "/CN=chromebox-gateway",
-    ], check=True)
-    os.chmod(KEY_FILE, 0o600)
-    os.chmod(CERT_FILE, 0o600)
-
-def main():
-    ap = argparse.ArgumentParser()
-    ap.add_argument("--port", type=int, default=8444)
-    ap.add_argument("--bind", default="100.123.153.75",
-                    help="tailnet IP; use 127.0.0.1 for local-only")
-    args = ap.parse_args()
-
-    ensure_cert()
-    context = ssl.SSLContext(ssl.PROTOCOL_TLS_SERVER)
-    context.load_cert_chain(CERT_FILE, KEY_FILE)
-
-    srv = HTTPServer((args.bind, args.port), Handler)
-    srv.socket = context.wrap_socket(srv.socket, server_side=True)
-    print("chromebox-gateway listening on https://%s:%d" % (args.bind, args.port),
-          file=sys.stderr)
-    print("ops: %s" % ", ".join(sorted(ALLOWLIST)), file=sys.stderr)
-    try:
-        srv.serve_forever()
-    except KeyboardInterrupt:
-        pass
-
-if __name__ == "__main__":
-    main()

bin/chromebox-watchdog.sh

@@ -1,225 +0,0 @@
-#!/usr/bin/env bash
-# chromebox-watchdog.sh [profile] — keep a chrome-box profile alive and healthy.
-# Checks: 1) chromium process for the profile is running,
-#         2) CDP responds and the chat page is present.
-# If unhealthy: kill any stale chrome for the profile and relaunch headless
-# via netvm-chrome.sh (profile dir persists session/cookies — the process is
-# disposable, the state is not). Mirrors operator-646's container
-# recover-after-rebuild.sh philosophy.
-# Runs every 2 min via systemd timer chromebox-watchdog-<profile>.timer.
-set -euo pipefail
-export XDG_RUNTIME_DIR="${XDG_RUNTIME_DIR:-/run/user/$(id -u)}"
-export DBUS_SESSION_BUS_ADDRESS="${DBUS_SESSION_BUS_ADDRESS:-unix:path=${XDG_RUNTIME_DIR}/bus}"
-# Prevent overlapping runs: the timer fires every 2 min but a relaunch
-# (kill + sleep 25 + chrome startup + page load) can exceed that, and two
-# concurrent runs kill each others chrome (observed 2026-10-03: pip flapped
-# with simultaneous "relaunch OK" and "relaunch FAILED").
-LOCK="/tmp/chromebox-watchdog-${1:-pip}.lock"
-# Tests source this file with CHROMEBOX_WATCHDOG_LIB_ONLY=1: they resolve
-# ports and call helpers without running checks, so no lock is needed.
-if [ "${CHROMEBOX_WATCHDOG_LIB_ONLY:-}" != "1" ]; then
-  exec 9>"$LOCK"
-  if ! flock -n 9; then
-    echo "[$(date -u +%FT%TZ)] [$1] another watchdog run in progress, skipping" >&2
-    exit 0
-  fi
-fi
-
-PROFILE="${1:-pip}"
-NETVM_BIN="/home/super/Projects/NetVM/bin"
-LOG="/home/super/Projects/NetVM/chromebox-watchdog.log"
-CHROME_LOG="/home/super/Projects/NetVM/chromebox-${PROFILE}.log"
-
-# Rotate a log file past 10MB (keep one generation)
-rotate_log() {
-  local f="$1"
-  [ -f "$f" ] || return 0
-  local sz
-  sz=$(stat -c%s "$f" 2>/dev/null || echo 0)
-  if [ "$sz" -gt 10485760 ]; then
-    mv -f "$f" "$f.1"
-    echo "[$(date -u +%FT%TZ)] [$PROFILE] log rotated" > "$f"
-  fi
-}
-rotate_log "$LOG"
-# CHROME_LOG rotation happens after PROFILE is set (see below)
-
-# Ports come from the fleet registry, not a hardcoded list: every active
-# node (def/dev included) gets supervision automatically. The old 4-profile
-# case left dev/def unsupervised — a dead Warp tunnel paged forever with
-# no auto-recovery (2026-10-06 dev outage).
-CDP_PORT="$("$NETVM_BIN/netvm-registry.py" "$PROFILE" 2>/dev/null)" || {
-  echo "unknown profile: $PROFILE" >&2
-  exit 1
-}
-[ -n "$CDP_PORT" ] || { echo "unknown profile: $PROFILE" >&2; exit 1; }
-rotate_log "$CHROME_LOG"
-
-log() { echo "[$(date -u +%FT%TZ)] [$PROFILE] $*" | tee -a "$LOG"; }
-
-cdp_list() {
-  "$NETVM_BIN/netvm-exec.sh" "$PROFILE" -- curl -s -m 8 "http://127.0.0.1:$CDP_PORT/json/list" 2>/dev/null
-}
-
-# HEALTH_FAIL_REASON is set by healthy() on failure: which stage broke.
-HEALTH_FAIL_REASON=""
-healthy() {
-  HEALTH_FAIL_REASON=""
-  pgrep -f "chromium.*profiles/${PROFILE}/" >/dev/null 2>&1 \
-    || { HEALTH_FAIL_REASON="no chromium process for profile"; return 1; }
-  local list
-  list="$(cdp_list)" \
-    || { HEALTH_FAIL_REASON="CDP unreachable on :$CDP_PORT"; return 1; }
-  [[ "$list" == *'"type": "page"'* ]] \
-    || { HEALTH_FAIL_REASON="CDP up but no page target in list"; return 1; }
-  [[ "$list" == *'"url": "https://muse.ai'* ]] \
-    || { HEALTH_FAIL_REASON="CDP up but not on muse.ai"; return 1; }
-  warp_egress_healthy || return 1
-  return 0
-}
-
-# Stage 5: Warp egress health (2026-10-04). A partitioned browser (tunnel down,
-# CDP green) passes stages 1-4 while being unable to reach muse.ai. Check the
-# WireGuard handshake age and probe egress from inside the node's netns.
-# Sets HEALTH_FAIL_REASON with a distinct "warp egress down" prefix so
-# root-cause analysis can distinguish partitions from Chromium crashes.
-warp_egress_healthy() {
-  local wg_out hs_line age_s code
-  wg_out="$(sudo -n ip netns exec "warp-${PROFILE}" wg show 2>/dev/null)" \
-    || { HEALTH_FAIL_REASON="warp egress down (wg show failed in warp-${PROFILE})"; return 1; }
-  hs_line="$(printf '%s\n' "$wg_out" | grep -i "latest handshake" | head -1)"
-  [ -n "$hs_line" ] \
-    || { HEALTH_FAIL_REASON="warp egress down (no WireGuard handshake in warp-${PROFILE})"; return 1; }
-  # "latest handshake: 1 minute, 41 seconds ago" -> total seconds
-  age_s="$(printf '%s\n' "$hs_line" | python3 -c '
-import sys, re
-s = sys.stdin.read()
-m = re.search(r"(\d+)\s*hour", s); h = int(m.group(1)) if m else 0
-m = re.search(r"(\d+)\s*minute", s); mi = int(m.group(1)) if m else 0
-m = re.search(r"(\d+)\s*second", s); se = int(m.group(1)) if m else 0
-print(h*3600 + mi*60 + se)
-' 2>/dev/null)"
-  { [ -n "$age_s" ] && [ "$age_s" -ge 0 ]; } 2>/dev/null \
-    || { HEALTH_FAIL_REASON="warp egress down (unparseable handshake: $hs_line)"; return 1; }
-  [ "$age_s" -le 180 ] \
-    || { HEALTH_FAIL_REASON="warp egress down (handshake ${age_s}s old in warp-${PROFILE})"; return 1; }
-  code="$(sudo -n ip netns exec "warp-${PROFILE}" curl -m 5 -s -o /dev/null -w "%{http_code}" "https://muse.ai/" 2>/dev/null)" \
-    || { HEALTH_FAIL_REASON="warp egress down (egress probe curl failed in warp-${PROFILE})"; return 1; }
-  # Any 2xx/3xx means we reached muse.ai infra ("/" 307-redirects to the
-  # auth flow). The probe tests egress connectivity, not page content.
-  case "$code" in
-    2*|3*) return 0 ;;
-    *) HEALTH_FAIL_REASON="warp egress down (egress probe HTTP $code in warp-${PROFILE})"; return 1 ;;
-  esac
-}
-
-# Allow sourcing for tests without running checks.
-if [ "${CHROMEBOX_WATCHDOG_LIB_ONLY:-}" = "1" ]; then
-  return 0 2>/dev/null || exit 0
-fi
-
-if healthy; then
-  # Auto-reconcile idle workers for healthy profiles
-  # DISABLED 2026-10-06 by operator-646: kpi auto-spawn ignores job schedule fields;
-  # find_pending_work_for_node returns the alphabetically-first definition every tick,
-  # re-spawning and re-noticing every ~2min (pip b01 BOX-AUTO-WORKER loop, 29+ copies).
-  # Watchdog health path untouched. Re-enable once the spawner is schedule-aware.
-  # python3 "$NETVM_BIN/super-cli.py" kpi auto-spawn --node "$PROFILE" >>"$LOG" 2>&1 || true
-  exit 0
-fi
-
-# Relaunch-loop guard (2026-10-04): if the main browser for this profile
-# launched <2 min ago it's probably still starting up (CDP not yet bound).
-# Relaunching now would kill a healthy-but-slow cold start via netvm-chrome.sh
-# and reset the startup clock every cycle (observed: opm piled up 5 chromiums
-# because the guard only protected the kill step, not the relaunch). Skip the
-# entire cycle instead.
-# NOTE: match only the main browser process (--remote-debugging-port present,
-# no --type= flag). Renderer/gpu children (--type=renderer etc.) start later
-# than the main process and must not satisfy this check.
-recent_pid=""
-for _pid in $(pgrep -f "chromium.*--remote-debugging-port=${CDP_PORT}([[:space:]]|$)" 2>/dev/null); do
-  # Skip child processes (renderer, gpu, etc.) — only the main browser counts
-  if ps -o args= -p "$_pid" 2>/dev/null | grep -q -- "--type="; then
-    continue
-  fi
-  _start=$(date -d "$(ps -o lstart= -p "$_pid" 2>/dev/null)" +%s 2>/dev/null || echo 0)
-  _now=$(date +%s)
-  if [ $(( _now - _start )) -lt 120 ] && [ "$_start" -gt 0 ]; then
-    recent_pid="$_pid"
-    break
-  fi
-done
-if [ -n "$recent_pid" ]; then
-  log "browser launched recently (pid $recent_pid), skipping relaunch (probably still starting)"
-  exit 0
-fi
-# Warp-partition recovery (2026-10-04): relaunching Chrome cannot fix a dead
-# Warp tunnel — the new browser would fail the same egress check and the
-# watchdog would loop. If the failure is warp-egress, restart the tunnel first
-# (netvm-node-up.sh is idempotent); only fall through to the Chrome relaunch
-# if the tunnel does not recover.
-case "$HEALTH_FAIL_REASON" in
-  "warp egress down"*)
-    log "warp partition detected ($HEALTH_FAIL_REASON), restarting tunnel via netvm-node-up.sh"
-    sudo -n "$NETVM_BIN/netvm-node-up.sh" "$PROFILE" >>"$LOG" 2>&1 || true
-    sleep 5
-    if healthy; then
-      log "tunnel restart recovered warp egress, chrome relaunch not needed"
-      exit 0
-    fi
-    log "tunnel restart did not recover egress ($HEALTH_FAIL_REASON), proceeding with chrome relaunch"
-    ;;
-esac
-log "unhealthy ($HEALTH_FAIL_REASON), relaunching chromebox"
-# bracket trick so pkill never matches its own command line
-pat="profiles/${PROFILE:0:${#PROFILE}-1}[${PROFILE: -1}]/"
-pkill -f "chromium.*$pat" 2>/dev/null || true
-sleep 3
-# Clean up stale singleton symlinks that break subsequent browser startup
-rm -f "/home/super/.local/share/chrome-box/profiles/${PROFILE}/home/.config/chromium/SingletonLock" \
-      "/home/super/.local/share/chrome-box/profiles/${PROFILE}/home/.config/chromium/SingletonSocket" \
-      "/home/super/.local/share/chrome-box/profiles/${PROFILE}/home/.config/chromium/SingletonCookie" 2>/dev/null || true
-for _sc in $(systemctl --user list-units --type=scope --plain --no-legend 2>/dev/null | awk '{print $1}' | grep -E "^netvm-chrome-${PROFILE}-"); do
-  systemctl --user stop "$_sc" 2>/dev/null || true
-done
-
-# Relaunch in its own systemd scope so it survives this oneshot run.
-# nohup/setsid do NOT escape: this timer's service uses KillMode=control-group
-# and systemd SIGKILLs everything in the cgroup at teardown (observed
-# 2026-10-03: every relaunch "recovered" then died seconds later). A transient
-# scope escapes the service cgroup; the scoped process inherits these fds so
-# the log redirect below still captures chromium's output.
-# NOTE: systemd-run --scope WAITS for the scope's processes (even --no-block,
-# verified 2026-10-03), so it must be backgrounded — the scope is an
-# independent unit and outlives the wrapper.
-# NOTE: --cdp-port is pinned explicitly. netvm-chrome.sh defaults to a
-# hash-derived port (9222+...) which will NOT match the registry port that
-# muse-chat-api.py uses — a relaunch on the wrong port looks healthy to the
-# launcher but is unreachable to the API (observed 2026-10-03: pip relaunched
-# on 9278 instead of 9420, watchdog looped on "relaunch FAILED").
-# 9>&-: do NOT let the backgrounded launcher inherit the watchdog lock fd.
-# Inherited flock fds wedge the lock forever (observed 2026-10-04: opm/muse
-# launchers held their profile lock for 100+ min, every later watchdog run
-# skipped as "another run in progress" — the watchdog was silently dead).
-systemd-run --user --scope --unit="netvm-chrome-${PROFILE}-$(date +%s)" \
-  "$NETVM_BIN/netvm-chrome.sh" --headless --cdp-port "$CDP_PORT" "$PROFILE" "https://muse.ai" \
-  >>"$CHROME_LOG" 2>&1 < /dev/null 9>&- &
-# Retry with backoff (2026-10-04): cold starts (fresh egress IP, Cloudflare
-# handshake) can take >25s for the page title to appear. A single check after
-# 25s kills working-but-slow browsers. Try up to 4 times, 15s apart (~60s
-# window), logging each attempt. Only declare FAILED if all attempts fail.
-_relaunch_ok=0
-for _attempt in 1 2 3 4; do
-  sleep 15
-  if healthy; then
-    log "relaunch OK (attempt $_attempt)"
-    _relaunch_ok=1
-    break
-  fi
-  log "relaunch attempt $_attempt not healthy yet ($HEALTH_FAIL_REASON)"
-done
-if [ "$_relaunch_ok" -ne 1 ]; then
-  log "relaunch FAILED ($HEALTH_FAIL_REASON) \u2014 needs operator attention"
-  exit 1
-fi

bin/completion-audit.py

@@ -1,312 +0,0 @@
-#!/usr/bin/env python3
-"""Completion auditor: prove work gets done, or say exactly where it stalls.
-
-Runs on a 15-minute systemd timer (systemd/completion-audit.*). Reads
-job-log.jsonl, swarms.json, and followups.json; computes the completion
-funnel per job family plus swarm drain and followup backlog; writes a JSON
-report under logs/ and posts a compact digest to the ops heartbeat sidechat
-when degraded (or a heartbeat summary every 6h when green).
-
-Read-only except the digest DM and its own log/state files. Exit 0 always
-on a completed audit; tracebacks (real errors) fail the timer visibly.
-"""
-import argparse
-import json
-import os
-import re
-import subprocess
-import sys
-from collections import Counter, defaultdict
-from datetime import datetime, timezone, timedelta
-from pathlib import Path
-
-REPO_ROOT = Path(__file__).resolve().parent.parent
-JOB_LOG = REPO_ROOT / "job-log.jsonl"
-SWARMS_FILE = REPO_ROOT / "swarms.json"
-FOLLOWUPS_FILE = REPO_ROOT / "followups.json"
-LOGS_DIR = REPO_ROOT / "logs"
-STATE_FILE = LOGS_DIR / "completion-audit-state.json"
-
-_JOB_ID_RE = re.compile(r"^(.+)-(\d{8})-(\d{6})-([0-9a-f]{8})$")
-HEARTBEAT_INTERVAL_H = 6
-STALE_RUNNING_MIN = 90
-SILENT_MIN_SENT = 3
-
-
-def utcnow():
-    return datetime.now(timezone.utc)
-
-
-def family_of(job_id):
-    """Strip the dispatch suffix (<name>-YYYYMMDD-HHMMSS-<hex8>) to the family."""
-    m = _JOB_ID_RE.match(job_id or "")
-    return m.group(1) if m else (job_id or "?")
-
-
-def parse_ts(ts):
-    try:
-        t = datetime.fromisoformat(str(ts))
-    except Exception:
-        return None
-    if t.tzinfo is None:
-        t = t.replace(tzinfo=timezone.utc)
-    return t
-
-
-def compute_funnel(events, cutoff):
-    """Aggregate job-log events since cutoff.
-
-    Returns (families, tools) where families maps family -> counters and
-    tools holds global tool_exec stats. Pure over the event list.
-    """
-    families = defaultdict(lambda: Counter())
-    tools = Counter()
-    tool_errs = Counter()
-    for e in events:
-        t = parse_ts(e.get("ts"))
-        if t is None or t < cutoff:
-            continue
-        ty = e.get("type")
-        if ty == "job_sent":
-            families[family_of(e.get("job_id"))]["sent"] += 1
-        elif ty == "job_dispatched":
-            families[family_of(e.get("job_id"))]["dispatched"] += 1
-        elif ty == "tool_exec":
-            tools["total"] += 1
-            if e.get("success"):
-                tools["ok"] += 1
-            else:
-                tools["fail"] += 1
-                tool_errs[e.get("op", "?")] += 1
-        elif ty == "job_result":
-            fam = family_of(e.get("job_id"))
-            families[fam]["results"] += 1
-            families[fam]["ok" if e.get("success") else "fail"] += 1
-        elif ty == "job_failed":
-            families[family_of(e.get("job_id"))]["failed"] += 1
-        elif ty == "fallback_executed":
-            families[family_of(e.get("job_id"))]["fallback_ok"] += 1
-        elif ty == "fallback_failed":
-            families[family_of(e.get("job_id"))]["fallback_fail"] += 1
-        elif ty == "proof_requested":
-            families[family_of(e.get("job_id"))]["proofs"] += 1
-    return families, {"tools": tools, "tool_errs": tool_errs}
-
-
-def swarm_drain(now):
-    """Status counts + stale-running slots from swarms.json."""
-    try:
-        data = json.load(open(SWARMS_FILE))
-    except Exception:
-        return {"error": "swarms.json unreadable"}, []
-    values = data.values() if isinstance(data, dict) else data
-    status = Counter()
-    stale = []
-    for s in values:
-        if not isinstance(s, dict):
-            continue
-        for sl in s.get("slots", []) or []:
-            status[sl.get("status", "?")] += 1
-            if sl.get("status") == "running":
-                upd = parse_ts(sl.get("updated_ts"))
-                if upd and (now - upd) > timedelta(minutes=STALE_RUNNING_MIN):
-                    stale.append({
-                        "swarm": s.get("swarm_id"),
-                        "slot": sl.get("slot"),
-                        "agent": sl.get("agent_id"),
-                        "idle_min": int((now - upd).total_seconds() // 60),
-                    })
-    return {"slots": dict(status)}, stale
-
-
-def followup_backlog(now):
-    """Pending/overdue/escalated counts from followups.json."""
-    try:
-        data = json.load(open(FOLLOWUPS_FILE))
-    except Exception:
-        return {"error": "followups.json unreadable"}
-    values = data.values() if isinstance(data, dict) else data
-    out = Counter()
-    for r in values:
-        if not isinstance(r, dict):
-            continue
-        st = r.get("status", "?")
-        out[st] += 1
-        if st == "pending":
-            dl = parse_ts(r.get("deadline"))
-            if dl and dl < now:
-                out["overdue"] += 1
-    return dict(out)
-
-
-def build_report(hours):
-    now = utcnow()
-    cutoff = now - timedelta(hours=hours)
-    events = []
-    try:
-        with open(JOB_LOG) as f:
-            for line in f:
-                line = line.strip()
-                if not line:
-                    continue
-                try:
-                    events.append(json.loads(line))
-                except Exception:
-                    continue
-    except FileNotFoundError:
-        pass
-    families, tools = compute_funnel(events, cutoff)
-    fam = {k: dict(v) for k, v in sorted(families.items())}
-    swarm, stale = swarm_drain(now)
-    backlog = followup_backlog(now)
-    totals = Counter()
-    for v in fam.values():
-        for k, n in v.items():
-            totals[k] += n
-    silent = sorted(
-        k for k, v in fam.items()
-        if v.get("sent", 0) >= SILENT_MIN_SENT and v.get("results", 0) == 0)
-    degraded_reasons = []
-    if silent:
-        degraded_reasons.append(f"{len(silent)} silent families: {', '.join(silent[:5])}")
-    if totals.get("failed"):
-        degraded_reasons.append(f"{totals['failed']} job_failed")
-    if tools["tools"].get("fail"):
-        top = tools["tool_errs"].most_common(3)
-        degraded_reasons.append(
-            "tool errors: " + ", ".join(f"{op}x{n}" for op, n in top))
-    if totals.get("fallback_fail"):
-        degraded_reasons.append(f"{totals['fallback_fail']} fallback_failed")
-    if stale:
-        degraded_reasons.append(f"{len(stale)} running slots idle >{STALE_RUNNING_MIN}m")
-    if backlog.get("overdue"):
-        degraded_reasons.append(f"{backlog['overdue']} overdue followups")
-    if backlog.get("escalated"):
-        degraded_reasons.append(f"{backlog['escalated']} escalated followups")
-    return {
-        "ts": now.isoformat(),
-        "window_h": hours,
-        "totals": dict(totals),
-        "tools": {k: dict(v) if isinstance(v, Counter) else v
-                  for k, v in tools.items()},
-        "families": fam,
-        "silent_families": silent,
-        "swarms": swarm,
-        "stale_running": stale[:10],
-        "followups": backlog,
-        "degraded": bool(degraded_reasons),
-        "reasons": degraded_reasons,
-    }
-
-
-def render_digest(rep):
-    t = rep["totals"]
-    tools = rep["tools"].get("tools", {})
-    lines = [
-        f"Completion audit ({rep['window_h']}h, {rep['ts'][:16]}Z)",
-        f"funnel: {t.get('sent', 0)} sent / {t.get('dispatched', 0)} dispatched / "
-        f"{tools.get('total', 0)} tool_exec / {t.get('results', 0)} results "
-        f"({t.get('ok', 0)} ok)",
-    ]
-    if rep["silent_families"]:
-        lines.append("silent: " + ", ".join(rep["silent_families"][:6]))
-    bits = []
-    if t.get("failed"):
-        bits.append(f"{t['failed']} job_failed")
-    if tools.get("fail"):
-        bits.append(f"{tools['fail']} tool errors")
-    if t.get("fallback_ok") or t.get("fallback_fail"):
-        bits.append(f"fallback {t.get('fallback_ok', 0)} ok / {t.get('fallback_fail', 0)} fail")
-    if t.get("proofs"):
-        bits.append(f"{t['proofs']} proof reqs")
-    if bits:
-        lines.append("flags: " + ", ".join(bits))
-    sw = rep["swarms"].get("slots", {})
-    if sw:
-        lines.append("swarms now: " + " / ".join(f"{v} {k}" for k, v in sorted(sw.items())))
-    if rep["stale_running"]:
-        lines.append(f"stale running: {len(rep['stale_running'])} slots (see report)")
-    fb = rep["followups"]
-    if fb and "error" not in fb:
-        lines.append(
-            f"followups now: {fb.get('pending', 0)} pending / {fb.get('overdue', 0)} "
-            f"overdue / {fb.get('escalated', 0)} escalated")
-    if rep["degraded"]:
-        lines.append("verdict: DEGRADED — " + "; ".join(rep["reasons"][:3]))
-    else:
-        lines.append("verdict: HEALTHY — work flowing, results landing")
-    return "\n".join(lines)
-
-
-def should_post(report):
-    """Post on degraded, else heartbeat at most every HEARTBEAT_INTERVAL_H."""
-    if report["degraded"]:
-        return True, "degraded"
-    try:
-        state = json.load(open(STATE_FILE))
-        last = parse_ts(state.get("last_heartbeat"))
-    except Exception:
-        last = None
-    if last is None or (utcnow() - last) > timedelta(hours=HEARTBEAT_INTERVAL_H):
-        return True, "heartbeat"
-    return False, "green-quiet"
-
-
-def post_digest(digest):
-    argv = [sys.executable, str(REPO_ROOT / "bin" / "dm.py"), "send",
-           "--agent", "super", "--to", "opm", "--target", "heartbeat", digest]
-    p = subprocess.run(argv, capture_output=True, text=True, timeout=120)
-    return p.returncode == 0, (p.stdout or p.stderr or "").strip()[:300]
-
-
-def save_report(report):
-    LOGS_DIR.mkdir(parents=True, exist_ok=True)
-    stamp = report["ts"].replace("+00:00", "Z").replace(":", "")
-    dated = LOGS_DIR / f"completion-audit-{stamp[:15]}.json"
-    body = json.dumps(report, indent=2)
-    dated.write_text(body, encoding="utf-8")
-    latest = LOGS_DIR / "completion-audit-latest.json"
-    tmp = LOGS_DIR / f".completion-audit-latest.tmp.{os.getpid()}"
-    tmp.write_text(body, encoding="utf-8")
-    os.replace(tmp, latest)
-    return dated
-
-
-def main():
-    ap = argparse.ArgumentParser(description="Completion funnel auditor")
-    ap.add_argument("--hours", type=int, default=24)
-    ap.add_argument("--post", dest="post", action="store_true", default=True)
-    ap.add_argument("--no-post", dest="post", action="store_false")
-    ap.add_argument("--json", action="store_true", help="Print raw report JSON")
-    args = ap.parse_args()
-
-    report = build_report(args.hours)
-    path = save_report(report)
-    if args.json:
-        print(json.dumps(report, indent=2))
-    else:
-        print(render_digest(report))
-    print(f"\nreport: {path}")
-    if not args.post:
-        print("post: skipped (--no-post)")
-        return 0
-    do_post, why = should_post(report)
-    if not do_post:
-        print(f"post: skipped ({why})")
-        return 0
-    ok, detail = post_digest(render_digest(report))
-    print(f"post: {'delivered' if ok else 'FAILED'} ({why}) {detail[:120]}")
-    if ok and why == "heartbeat":
-        try:
-            state = {}
-            if STATE_FILE.exists():
-                state = json.loads(STATE_FILE.read_text(encoding="utf-8"))
-            state["last_heartbeat"] = report["ts"]
-            STATE_FILE.write_text(json.dumps(state, indent=2), encoding="utf-8")
-        except Exception as e:
-            print(f"warning: state save failed: {e}")
-    return 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/cred-client.py

@@ -1,416 +0,0 @@
-#!/usr/bin/env python3
-"""cred-client.py — Agent API client & CLI for cred.muse-dev.online.
-
-Provides programmatic and CLI access for agents and operators to:
-- initiate: start onboarding for a client email/node
-- submit-otp: submit verification code transiently
-- status: query onboarding & vitality state for a node
-- list: list fleet accounts, emails, and statuses
-
-Authentication:
-- Machine identity signature via ~/.ssh/muse-health (machine bl)
-- Or Bearer token from CRED_TOKEN / OPERATOR_TOKEN environment variable.
-
-Usage:
-  cred-client.py initiate --node <node> --email <email> [--service muse] [--account-name <name>]
-  cred-client.py submit-otp --node <node> --otp <code> [--email <email>]
-  cred-client.py status --node <node> [--json]
-  cred-client.py list [--json]
-"""
-import argparse
-import datetime
-import importlib.util
-import json
-import os
-import re
-import subprocess
-import sys
-import tempfile
-import urllib.error
-import urllib.request
-
-NETVM_DIR = os.environ.get("NETVM_DIR", "/home/super/Projects/NetVM")
-KEY_DEFAULT = os.path.expanduser("~/.ssh/muse-health")
-CRED_API_DEFAULT = os.environ.get("CRED_API_URL", "https://cred.muse-dev.online/api/cred")
-
-
-def load_registry():
-    path = os.path.join(NETVM_DIR, "bin", "netvm-registry.py")
-    try:
-        spec = importlib.util.spec_from_file_location("netvm_registry", path)
-        mod = importlib.util.module_from_spec(spec)
-        spec.loader.exec_module(mod)
-        return mod
-    except Exception:
-        return None
-
-
-def get_accounts():
-    """Parse ACCOUNTS.md into a structured list of accounts."""
-    path = os.path.join(NETVM_DIR, "ACCOUNTS.md")
-    accounts = []
-    if not os.path.exists(path):
-        return accounts
-    with open(path) as f:
-        for line in f:
-            line = line.strip()
-            if not line.startswith("|") or line.startswith("| agent") or line.startswith("|-------"):
-                continue
-            cols = [c.strip() for c in line.strip("|").split("|")]
-            if len(cols) >= 9:
-                accounts.append({
-                    "agent": cols[0],
-                    "node": cols[1],
-                    "profile": cols[2],
-                    "login_type": cols[3],
-                    "meta_label": cols[4],
-                    "email": cols[5],
-                    "phone_otp": cols[6],
-                    "instagram_linked": cols[7],
-                    "status": cols[8],
-                    "egress_ip": cols[9] if len(cols) > 9 else "",
-                    "cdp_port": cols[10] if len(cols) > 10 else "",
-                    "display_name": cols[11] if len(cols) > 11 else "",
-                    "notes": cols[12] if len(cols) > 12 else ""
-                })
-    return accounts
-
-
-def sign_payload(payload_bytes, key_path=KEY_DEFAULT, namespace="cred"):
-    """Sign payload using SSH ed25519 key (zero secret across wire)."""
-    if not os.path.exists(key_path):
-        return None
-    tmp = tempfile.mkdtemp()
-    try:
-        data_file = os.path.join(tmp, "data")
-        with open(data_file, "wb") as f:
-            f.write(payload_bytes)
-        r = subprocess.run(
-            ["ssh-keygen", "-Y", "sign", "-f", key_path, "-n", namespace, data_file],
-            capture_output=True, text=True
-        )
-        if r.returncode != 0:
-            return None
-        with open(data_file + ".sig") as f:
-            return f.read().strip()
-    finally:
-        subprocess.run(["rm", "-rf", tmp], capture_output=True)
-
-
-class CredClient:
-    def __init__(self, api_url=CRED_API_DEFAULT, key_path=KEY_DEFAULT):
-        self.api_url = api_url.rstrip("/")
-        self.key_path = key_path
-        self.token = os.environ.get("CRED_TOKEN") or os.environ.get("OPERATOR_TOKEN")
-
-    def run_local_driver(self, node, step, email, otp=None, account_name=None):
-        """Execute local onboard-driver.py inside the node's netns."""
-        exec_script = os.path.join(NETVM_DIR, "bin", "netvm-exec.sh")
-        driver_script = os.path.join(NETVM_DIR, "bin", "onboard-driver.py")
-        cmd = [exec_script, node, "--", sys.executable, driver_script,
-               "--node", node, "--service", "muse", "--id-type", "email", "--step", step]
-        if account_name:
-            cmd += ["--account-name", account_name]
-
-        input_data = email + "\n"
-        if otp:
-            input_data += str(otp) + "\n"
-
-        p = subprocess.run(cmd, input=input_data, capture_output=True, text=True, timeout=240)
-        return p.returncode, p.stdout.strip(), p.stderr.strip()
-
-    def initiate(self, node, email, service="muse", account_name=None):
-        """Initiate client onboarding."""
-        ret, stdout, stderr = self.run_local_driver(node, "initiate", email, account_name=account_name)
-        if ret == 0:
-            return {"status": "active", "node": node, "email": email, "message": "Already authenticated and session active."}
-        elif ret == 2:
-            return {"status": "awaiting_otp", "node": node, "email": email, "message": "OTP verification code sent. Awaiting input."}
-        elif ret == 3:
-            return {"status": "needs_human", "node": node, "email": email, "message": "Multiple accounts match identifier. Manual selection required."}
-        else:
-            return {"status": "error", "node": node, "email": email, "code": ret, "detail": stderr or stdout}
-
-    def submit_otp(self, node, otp, email=None, service="muse"):
-        """Submit transient verification code."""
-        if not email:
-            # Look up email from ACCOUNTS.md
-            for acct in get_accounts():
-                if acct["node"] == node and acct["email"] not in ("-", ""):
-                    email = acct["email"]
-                    break
-            if not email:
-                email = "unknown"
-
-        ret, stdout, stderr = self.run_local_driver(node, "submit", email, otp=otp)
-        if ret == 0:
-            return {"status": "active", "node": node, "email": email, "message": "Authentication successful. Chat session active."}
-        elif ret == 3:
-            return {"status": "needs_human", "node": node, "email": email, "message": "Multiple accounts match identifier."}
-        else:
-            return {"status": "error", "node": node, "email": email, "code": ret, "detail": stderr or stdout}
-
-    def status(self, node):
-        """Check status of a node."""
-        accounts = get_accounts()
-        target = next((a for a in accounts if a["node"] == node), None)
-        port = None
-        reg = load_registry()
-        if reg:
-            port = reg.port_for(node)
-
-        # Vitality check
-        alive = False
-        detail = ""
-        if port:
-            try:
-                checker = os.path.join(NETVM_DIR, "bin", "accounts-health.py")
-                cmd = ["sudo", "-n", "ip", "netns", "exec", f"warp-{node}", sys.executable, checker, str(port)]
-                r = subprocess.run(cmd, capture_output=True, text=True, timeout=10)
-                if r.returncode == 0:
-                    data = json.loads(r.stdout.strip().splitlines()[-1])
-                    alive = data.get("session_alive", False)
-                    detail = data.get("detail", "")
-            except Exception as e:
-                detail = str(e)
-
-        return {
-            "node": node,
-            "status": target["status"] if target else "unregistered",
-            "email": target["email"] if target else "-",
-            "cdp_port": port,
-            "session_alive": alive,
-            "detail": detail
-        }
-
-    def list_all(self):
-        """List all accounts and live statuses."""
-        res = []
-        for a in get_accounts():
-            st = self.status(a["node"])
-            res.append(st)
-        return res
-
-
-    def notify_operator(self, subject, body, to_email="defnotabotnet@gmail.com"):
-        """Send notification to operator via local MTA (msmtp or mail)."""
-        sent = False
-        err = None
-        # Try msmtp first
-        try:
-            p = subprocess.Popen(["msmtp", to_email], stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
-            msg = f"Subject: {subject}\nTo: {to_email}\nFrom: NetVM Automation <root@bl>\n\n{body}\n"
-            stdout, stderr = p.communicate(input=msg)
-            if p.returncode == 0:
-                sent = True
-            else:
-                err = stderr.strip() or stdout.strip()
-        except Exception as e:
-            err = str(e)
-
-        if not sent:
-            # Fallback to mail / s-nail
-            try:
-                p = subprocess.Popen(["mail", "-s", subject, to_email], stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
-                stdout, stderr = p.communicate(input=body)
-                if p.returncode == 0:
-                    sent = True
-                else:
-                    err = stderr.strip() or stdout.strip()
-            except Exception as e:
-                err = str(e)
-
-        return {"sent": sent, "recipient": to_email, "error": err if not sent else None}
-
-    def link_instagram(self, node, notify=False):
-        """Fetch the Meta Accounts Center OAuth URL and provide one-tap Tailscale link."""
-        portal_script = os.path.join(NETVM_DIR, "bin", "tailscale-verify-portal.py")
-        ts_ip = "100.123.153.75"
-        ts_dns = "bl.tailfb5960.ts.net"
-        try:
-            import importlib.util
-            spec = importlib.util.spec_from_file_location("portal", portal_script)
-            portal_mod = importlib.util.module_from_spec(spec)
-            spec.loader.exec_module(portal_mod)
-            ts_ip = portal_mod.get_tailscale_ip()
-            ts_dns = portal_mod.get_tailscale_dns()
-            ig_data = portal_mod.fetch_node_ig_link(node)
-        except Exception as e:
-            ig_data = {"error": str(e)}
-
-        direct_url = ig_data.get("url") if isinstance(ig_data, dict) else None
-        portal_url = f"http://{ts_dns}:8765/verify/{node}"
-        portal_ip_url = f"http://{ts_ip}:8765/verify/{node}"
-
-        email_result = None
-        if notify and direct_url:
-            subject = f"[NetVM Action Required] Link Instagram for Node '{node}'"
-            body = (
-                f"Node '{node}' is waiting at the Muse age verification gate.\n\n"
-                f"Tap the Tailscale portal link from your device to approve:\n"
-                f"  {portal_url}\n"
-                f"  (or {portal_ip_url})\n\n"
-                f"Direct OAuth URL:\n"
-                f"  {direct_url}\n\n"
-                f"After approving in Instagram, NetVM will automatically transition node '{node}' to active."
-            )
-            email_result = self.notify_operator(subject, body)
-
-        return {
-            "node": node,
-            "status": "needs_verification",
-            "portal_url": portal_url,
-            "portal_ip_url": portal_ip_url,
-            "direct_oauth_url": direct_url,
-            "error": ig_data.get("error") if isinstance(ig_data, dict) else None,
-            "email_notified": email_result.get("sent") if email_result else False
-        }
-
-    def audit_meta(self, node):
-        """Query Meta Accounts Center for linked profiles and security status."""
-        meta_script = os.path.join(NETVM_DIR, "bin", "meta-acct.py")
-        cmd = ["sudo", "-n", "ip", "netns", "exec", f"warp-{node}", sys.executable, meta_script, "list-linked", node]
-        try:
-            res = subprocess.run(cmd, capture_output=True, text=True, timeout=15)
-            out = res.stdout.strip()
-            if out:
-                # Find outermost JSON object
-                start = out.find("{")
-                end = out.rfind("}")
-                if start >= 0 and end >= start:
-                    return json.loads(out[start:end+1])
-                return json.loads(out)
-            return {"error": res.stderr.strip() or "No output from meta-acct.py"}
-        except Exception as e:
-            return {"error": str(e)}
-
-
-def main():
-    common = argparse.ArgumentParser(add_help=False)
-    common.add_argument("--json", action="store_true", help="Output JSON response")
-
-    p = argparse.ArgumentParser(description="cred-client: Agent API client for cred onboarding", parents=[common])
-    sub = p.add_subparsers(dest="command", required=True)
-
-    # initiate
-    p_init = sub.add_parser("initiate", parents=[common], help="Initiate onboarding flow for a node")
-    p_init.add_argument("--node", required=True, help="Node label (e.g. opm, pip, client1)")
-    p_init.add_argument("--email", required=True, help="Client login email")
-    p_init.add_argument("--service", default="muse", help="Service name (default: muse)")
-    p_init.add_argument("--account-name", default=None, help="Display name hint for multi-account selector")
-
-    # submit-otp
-    p_otp = sub.add_parser("submit-otp", parents=[common], help="Submit transient OTP verification code")
-    p_otp.add_argument("--node", required=True, help="Node label")
-    p_otp.add_argument("--otp", required=True, help="6-digit verification code")
-    p_otp.add_argument("--email", default=None, help="Client email (optional, auto-detected from registry)")
-
-    # status
-    p_stat = sub.add_parser("status", parents=[common], help="Query onboarding and session status for a node")
-    p_stat.add_argument("--node", required=True, help="Node label")
-
-    # link-instagram
-    p_link = sub.add_parser("link-instagram", parents=[common], help="Generate Tailscale one-tap portal & OAuth link for age verification")
-    p_link.add_argument("--node", required=True, help="Node label")
-    p_link.add_argument("--notify", action="store_true", help="Send email alert to operator via local MTA")
-
-    # meta-audit
-    p_meta = sub.add_parser("meta-audit", parents=[common], help="Query Meta Accounts Center for linked profiles and security status")
-    p_meta.add_argument("--node", required=True, help="Node label")
-
-    # list
-    sub.add_parser("list", parents=[common], help="List all registered nodes and vitality statuses")
-
-    args = p.parse_args()
-    client = CredClient()
-
-    if args.command == "meta-audit":
-        res = client.audit_meta(args.node)
-        if args.json:
-            print(json.dumps(res, indent=2))
-        else:
-            print(f"\n=== META ACCOUNTS CENTER AUDIT: {args.node} ===")
-            if res.get("error"):
-                print(f"Error: {res['error']}", file=sys.stderr)
-                sys.exit(1)
-            print(f"Meta Account Email: {res.get('email') or '(none / phone-only)'}")
-            profiles = res.get("profiles", [])
-            print(f"Linked Profiles ({len(profiles)}):")
-            for p_info in profiles:
-                print(f"  - [{p_info.get('type')}] {p_info.get('name')}")
-            sys.exit(0)
-
-    if args.command == "link-instagram":
-        res = client.link_instagram(args.node, notify=args.notify)
-        if args.json:
-            print(json.dumps(res, indent=2))
-        else:
-            print(f"\n=== INSTAGRAM LINKING: {args.node} ===")
-            if res.get("error"):
-                print(f"Error: {res['error']}", file=sys.stderr)
-                sys.exit(1)
-            print(f"Tailscale Portal: {res['portal_url']}")
-            print(f"Direct IP Portal: {res['portal_ip_url']}")
-            print(f"OAuth URL:        {res['direct_oauth_url'][:80]}...")
-            if args.notify:
-                print(f"Email Notified:   {'YES' if res['email_notified'] else 'FAILED'}")
-            print("\nTap either Tailscale link from your phone/browser to complete Meta age verification.")
-            sys.exit(0)
-
-    if args.command == "initiate":
-        res = client.initiate(args.node, args.email, service=args.service, account_name=args.account_name)
-        if args.json:
-            print(json.dumps(res, indent=2))
-        else:
-            st = res.get("status")
-            if st == "awaiting_otp":
-                print(f"[APPROVAL_NEEDED] Code sent to [redacted] for node '{args.node}'.")
-                print(f"Submit OTP with: super cred submit-otp --node {args.node} --otp <code>")
-                sys.exit(2)
-            elif st == "active":
-                print(f"[SUCCESS] Node '{args.node}' is already authenticated and active.")
-                sys.exit(0)
-            else:
-                print(f"[{st.upper()}] {res.get('message') or res.get('detail')}")
-                sys.exit(res.get("code", 1))
-
-    elif args.command == "submit-otp":
-        res = client.submit_otp(args.node, args.otp, email=args.email)
-        if args.json:
-            print(json.dumps(res, indent=2))
-        else:
-            st = res.get("status")
-            if st == "active":
-                print(f"[SUCCESS] Node '{args.node}' successfully signed in!")
-                sys.exit(0)
-            else:
-                print(f"[{st.upper()}] {res.get('message') or res.get('detail')}")
-                sys.exit(res.get("code", 1))
-
-    elif args.command == "status":
-        res = client.status(args.node)
-        if args.json:
-            print(json.dumps(res, indent=2))
-        else:
-            print(f"Node:          {res['node']}")
-            print(f"Email:         {res['email']}")
-            print(f"Status:        {res['status']}")
-            print(f"CDP Port:      {res['cdp_port'] or '-'}")
-            print(f"Session Alive: {'YES' if res['session_alive'] else 'NO'}")
-            if res["detail"]:
-                print(f"Detail:        {res['detail']}")
-
-    elif args.command == "list":
-        items = client.list_all()
-        if args.json:
-            print(json.dumps(items, indent=2))
-        else:
-            print(f"{'NODE':<10} {'STATUS':<14} {'CDP':<6} {'ALIVE':<7} {'EMAIL'}")
-            print("-" * 65)
-            for it in items:
-                alive_str = "YES" if it["session_alive"] else "NO"
-                print(f"{it['node']:<10} {it['status']:<14} {str(it['cdp_port'] or '-'):<6} {alive_str:<7} {it['email']}")
-
-
-if __name__ == "__main__":
-    main()

bin/cred_client.py

@@ -1 +0,0 @@
-cred-client.py

bin/crypt-server.py

@@ -1,178 +0,0 @@
-#!/usr/bin/env python3
-"""
-crypt-server.py — Public cryptographic attestation and key directory for NetVM.
-
-Serves https://crypt.muse-dev.online/ (via cloudflared / reverse proxy).
-Endpoints:
-  GET /                         -> Service directory / health JSON
-  GET /health                   -> Health check
-  GET /keys/allowed_signers     -> OpenSSH allowed_signers formatted file
-  GET /keys/{identity}.pub      -> Individual public key
-  GET /proofs                   -> List known proof hashes / work order attestations
-  GET /proofs/{id}              -> Retrieve proof envelope and SSH signature
-  POST /proofs                  -> Submit / register a signed proof record
-"""
-
-import argparse
-import glob
-import json
-import os
-import re
-import ssl
-import sys
-from http.server import HTTPServer, BaseHTTPRequestHandler
-
-REPO_DIR = "/home/super/Projects/NetVM"
-SIGNERS_DIR = os.path.join(REPO_DIR, "dm-signers")
-PROOFS_DIR = os.path.join(REPO_DIR, "var", "proofs")
-
-os.makedirs(PROOFS_DIR, exist_ok=True)
-
-class CryptHandler(BaseHTTPRequestHandler):
-    server_version = "crypt-attestation/1.0"
-
-    def log_message(self, format, *args):
-        sys.stderr.write(f"crypt-server: {self.client_address[0]} - {format % args}\n")
-
-    def _send(self, code, content, content_type="application/json"):
-        if isinstance(content, (dict, list)):
-            body = json.dumps(content, indent=2).encode("utf-8")
-        elif isinstance(content, str):
-            body = content.encode("utf-8")
-        else:
-            body = bytes(content)
-
-        self.send_response(code)
-        self.send_header("Content-Type", content_type)
-        self.send_header("Content-Length", str(len(body)))
-        self.send_header("Access-Control-Allow-Origin", "*")
-        self.end_headers()
-        self.wfile.write(body)
-
-    def do_GET(self):
-        path = self.path.split("?")[0].rstrip("/")
-        if not path:
-            path = "/"
-
-        if path in ("/", "/health"):
-            self._send(200, {
-                "service": "crypt.muse-dev.online",
-                "status": "active",
-                "mode": "public_attestation",
-                "endpoints": [
-                    "/keys/allowed_signers",
-                    "/keys/<identity>.pub",
-                    "/proofs",
-                    "/proofs/<id>"
-                ]
-            })
-            return
-
-        if path == "/keys/allowed_signers":
-            allowed_path = os.path.join(SIGNERS_DIR, "allowed_signers")
-            if os.path.exists(allowed_path):
-                with open(allowed_path, "r") as f:
-                    data = f.read()
-                self._send(200, data, content_type="text/plain; charset=utf-8")
-            else:
-                self._send(404, {"error": "allowed_signers not found"})
-            return
-
-        m_key = re.match(r"^/keys/([a-zA-Z0-9_\-\.]+)\.pub$", path)
-        if m_key:
-            ident = m_key.group(1)
-            pub_path = os.path.join(SIGNERS_DIR, f"{ident}.pub")
-            if os.path.exists(pub_path):
-                with open(pub_path, "r") as f:
-                    data = f.read()
-                self._send(200, data, content_type="text/plain; charset=utf-8")
-            else:
-                self._send(404, {"error": f"Public key for {ident} not found"})
-            return
-
-        if path == "/proofs":
-            proof_files = glob.glob(os.path.join(PROOFS_DIR, "*.json"))
-            ids = [os.path.basename(p)[:-5] for p in proof_files]
-            self._send(200, {"proofs": sorted(ids)})
-            return
-
-        m_proof = re.match(r"^/proofs/([a-zA-Z0-9_\-]+)$", path)
-        if m_proof:
-            pid = m_proof.group(1)
-            pf = os.path.join(PROOFS_DIR, f"{pid}.json")
-            if os.path.exists(pf):
-                with open(pf, "r") as f:
-                    data = json.load(f)
-                self._send(200, data)
-            else:
-                self._send(404, {"error": f"Proof {pid} not found"})
-            return
-
-        self._send(404, {"error": "not found"})
-
-    def do_POST(self):
-        path = self.path.split("?")[0].rstrip("/")
-        if path == "/proofs":
-            try:
-                length = int(self.headers.get("Content-Length", 0))
-            except ValueError:
-                length = 0
-            if length <= 0 or length > 65536:
-                self._send(400, {"error": "invalid content length"})
-                return
-            try:
-                data = json.loads(self.rfile.read(length))
-            except Exception:
-                self._send(400, {"error": "malformed JSON"})
-                return
-
-            pid = data.get("id")
-            if not pid or not re.match(r"^[a-zA-Z0-9_\-]+$", pid):
-                self._send(400, {"error": "missing or invalid proof id"})
-                return
-
-            pf = os.path.join(PROOFS_DIR, f"{pid}.json")
-            with open(pf, "w") as f:
-                json.dump(data, f, indent=2)
-
-            self._send(201, {"status": "stored", "id": pid, "url": f"https://crypt.muse-dev.online/proofs/{pid}"})
-            return
-
-        self._send(404, {"error": "not found"})
-
-
-def main():
-    parser = argparse.ArgumentParser(description="Crypt attestation and key server")
-    parser.add_argument("--port", type=int, default=8446)
-    parser.add_argument("--host", default="100.123.153.75")
-    parser.add_argument("--ssl", action="store_true", help="Enable self-signed HTTPS")
-    args = parser.parse_args()
-
-    server = HTTPServer((args.host, args.port), CryptHandler)
-
-    if args.ssl:
-        cert_file = os.path.join(REPO_DIR, "ssl", "crypt-selfsigned.crt")
-        key_file = os.path.join(REPO_DIR, "ssl", "crypt-selfsigned.key")
-        os.makedirs(os.path.dirname(cert_file), exist_ok=True)
-        if not os.path.exists(cert_file):
-            import subprocess
-            subprocess.run([
-                "openssl", "req", "-x509", "-newkey", "rsa:2048",
-                "-keyout", key_file, "-out", cert_file,
-                "-days", "3650", "-nodes",
-                "-subj", "/CN=crypt.muse-dev.online"
-            ], check=True, capture_output=True)
-            os.chmod(key_file, 0o600)
-        ctx = ssl.SSLContext(ssl.PROTOCOL_TLS_SERVER)
-        ctx.load_cert_chain(cert_file, key_file)
-        server.socket = ctx.wrap_socket(server.socket, server_side=True)
-
-    print(f"Crypt server running on {'https' if args.ssl else 'http'}://{args.host}:{args.port}", file=sys.stderr)
-    try:
-        server.serve_forever()
-    except KeyboardInterrupt:
-        print("\nShutting down crypt server", file=sys.stderr)
-
-
-if __name__ == "__main__":
-    main()

bin/detect.py

@@ -1,266 +0,0 @@
-#!/usr/bin/env python3
-"""
-Side-chat to main-chat work siphon — detection rules (REPAIRED, agent 2 of 5).
-
-Fixes the false-positive ✅ COMPLETED relay at the source:
-  "Sending is disabled until this conversation can be verified." → COMPLETED
-  was caused by a SINGLE keyword ("verified") matching one regex.
-
-Repairs (see OUTPUT.md for rationale):
-  1. COMPLETED requires >= 2 DISTINCT pattern hits (weighted: the structured
-     `[RESULT ...] OK` marker counts 2 — it is the fleet's own machine-emitted
-     completion signal, far less ambiguous than a bare "done").
-  2. Negation guards: negation/failure-state words veto COMPLETED outright
-     (fail-closed: a negated completion claim is never relayed as complete).
-  3. Honest labeling: the fake "confidence 60%" (which literally meant "one
-     regex hit") is replaced by a keyword-hit count. SiphonHit.hits is the
-     authoritative field; `confidence` is kept for backward compatibility
-     but must NOT be rendered as a percentage anywhere user-facing.
-  4. Stale suppression: a message older than 15 minutes never relays as
-     COMPLETED. Pass message_ts (epoch seconds). monitor.py currently does
-     NOT pass a timestamp — agent 3 / the integrator must thread
-     message["ts"] through (see OUTPUT.md).
-
-DO NOT overwrite the original detect.py with this file until the integrator
-reconciles all 5 agents' outputs.
-"""
-
-import re
-import time
-from dataclasses import dataclass, field
-from typing import Optional
-
-
-@dataclass
-class SiphonHit:
-    category: str       # COMPLETED, BLOCKER, DECISION, ALERT, MILESTONE
-    confidence: float   # LEGACY — kept for API compatibility only.
-                        # Do NOT render as "confidence NN%"; it is not a
-                        # reliability measure. See `hits`.
-    hits: int = 0       # AUTHORITATIVE — distinct keyword-pattern hits
-                        # (weighted; see COMPLETED_PATTERN_WEIGHTS).
-    summary: str = ""   # one-line summary for main chat
-    thread_id: str = "" # source side chat
-    message_id: str = ""  # source message
-    author: str = ""    # INTEGRATOR (agent 3 absent) — the message's real
-                        # author, plumbed from message["author"] by
-                        # monitor.py. Empty = unknown; NEVER substitute the
-                        # thread's registered agent silently (see
-                        # format_siphon).
-    # Full message text is NOT stored here — main chat gets a summary
-    # plus a link back, never the full content (safety: no sensitive
-    # data siphoned verbatim).
-
-
-# --- Keyword sets (configurable) ---
-
-COMPLETED_PATTERNS = [
-    re.compile(r'\b(done|completed|finished|deployed|shipped|live|verified)\b', re.I),
-    re.compile(r'\b(all|tests?)\s+(pass|green|passing)\b', re.I),
-    re.compile(r'\[RESULT[^\]]*\]\s*OK', re.I),
-    re.compile(r'\b(merged|committed|pushed|published)\b', re.I),
-]
-
-# Weighted hits: the structured [RESULT] OK marker is the fleet's own
-# machine-emitted completion signal — unambiguous enough to stand alone.
-COMPLETED_PATTERN_WEIGHTS = {0: 1, 1: 1, 2: 2, 3: 1}
-COMPLETED_MIN_WEIGHT = 2  # >= 2 distinct pattern hits (or one [RESULT] OK)
-
-BLOCKER_PATTERNS = [
-    re.compile(r'\b(blocked|stuck|failing|broken|down|error|failed)\b', re.I),
-    re.compile(r'\b(need|needs|waiting)\s+(your|approval|input|decision)\b', re.I),
-    re.compile(r'\b(can\'t|cannot|unable to)\b', re.I),
-    re.compile(r'\[RESULT[^\]]*\]\s*(FAIL|ERROR)', re.I),
-]
-
-DECISION_PATTERNS = [
-    re.compile(r'\b(should (i|we)|shall i|want me to)\b', re.I),
-    re.compile(r'\b(your call|needs? your|awaiting your)\b', re.I),
-    re.compile(r'\b(approve|approval)\b.*\?', re.I),
-    re.compile(r'^(yes|no)\s*\?\s*$', re.I),
-]
-
-ALERT_PATTERNS = [
-    re.compile(r'\b(security|vulnerability|breach|compromised|exploit)\b', re.I),
-    re.compile(r'\b(urgent|critical|emergency|asap)\b', re.I),
-    re.compile(r'\b(502|503|500)\b.*\b(error|down)\b', re.I),
-    re.compile(r'\b(ssh|tunnel).*\b(down|broken|failed)\b', re.I),
-]
-
-MILESTONE_PATTERNS = [
-    re.compile(r'\b(milestone|phase \d+ (complete|done)|shipped v)\b', re.I),
-    re.compile(r'\b(all \d+ (items? )?done)\b', re.I),
-]
-
-# Patterns that suppress siphoning (safety)
-SUPPRESS_PATTERNS = [
-    re.compile(r'\b(password|secret|token|key|pin)\s*[:=]', re.I),
-    re.compile(r'-----BEGIN', re.I),  # never siphon key material
-    re.compile(r'\[do not siphon\]', re.I),  # explicit opt-out marker
-]
-
-# --- Negation guards: any match vetoes COMPLETED (fail-closed) ---
-# A completion claim in the presence of negation / failure-state language
-# is never relayed as ✅ COMPLETED, no matter how many keywords hit.
-NEGATION_GUARDS = [
-    # explicit negation
-    re.compile(r'\b(not|never|no|nothing|none|neither|nor)\b', re.I),
-    re.compile(r"\b(do not|don't|didn't|doesn't|won't|can't|cannot|isn't|aren't|"
-               r"wasn't|weren't|haven't|hasn't|hadn't|couldn't|shouldn't)\b", re.I),
-    # incompleteness hedges
-    re.compile(r'\b(still|yet|pending|unfinished|incomplete)\b', re.I),
-    # failure-state words (a "completed" message containing these is suspect)
-    re.compile(r'\b(broken|failed|failing|failure|down|stuck|blocked|disabled|'
-               r'error|errors|crash|crashed)\b', re.I),
-    # hedging conjunctions ("deployed, but tests are red")
-    re.compile(r'\b(but|however|although|though)\b', re.I),
-]
-
-# Messages older than this never relay as COMPLETED (seconds).
-COMPLETED_MAX_AGE_S = 15 * 60
-
-
-def _match_score(text: str, patterns) -> float:
-    """Legacy confidence for non-COMPLETED categories (unchanged)."""
-    hits = sum(1 for p in patterns if p.search(text))
-    if hits == 0:
-        return 0.0
-    # Diminishing returns: 1 hit = 0.6, 2 = 0.8, 3+ = 0.95
-    return min(0.95, 0.6 + (hits - 1) * 0.2)
-
-
-def _completed_weight(text: str):
-    """
-    Return (weighted_hits, distinct_hits, matched_pattern_indexes) for
-    COMPLETED_PATTERNS. Weighted: [RESULT] OK counts 2.
-    """
-    matched = [i for i, p in enumerate(COMPLETED_PATTERNS) if p.search(text)]
-    weight = sum(COMPLETED_PATTERN_WEIGHTS.get(i, 1) for i in matched)
-    return weight, len(matched), matched
-
-
-def _is_negated(text: str) -> bool:
-    """True if any negation guard fires anywhere in the text."""
-    return any(p.search(text) for p in NEGATION_GUARDS)
-
-
-def _extract_summary(text: str, max_len: int = 120) -> str:
-    """Extract a safe one-line summary. Strips to first meaningful line."""
-    # Take first non-empty line, truncate
-    for line in text.strip().split('\n'):
-        line = line.strip()
-        if line and len(line) > 10:
-            if len(line) > max_len:
-                return line[:max_len - 3] + '...'
-            return line
-    return text[:max_len]
-
-
-def detect(text: str, thread_id: str, message_id: str,
-           min_confidence: float = 0.6,
-           message_ts: Optional[float] = None) -> Optional[SiphonHit]:
-    """
-    Check a side chat message for siphon-worthy content.
-    Returns SiphonHit or None.
-
-    message_ts: epoch seconds of the original message (optional). Messages
-    older than COMPLETED_MAX_AGE_S (15 min) never relay as COMPLETED.
-    NOTE: monitor.py does not currently pass a timestamp — agent 3 / the
-    integrator must thread message["ts"] through the detect() call.
-    """
-    # Safety: suppress sensitive content
-    for p in SUPPRESS_PATTERNS:
-        if p.search(text):
-            return None
-
-    # Stale suppression applies to COMPLETED only.
-    completed_allowed = True
-    if message_ts is not None:
-        try:
-            age = time.time() - float(message_ts)
-            if age > COMPLETED_MAX_AGE_S:
-                completed_allowed = False
-        except (TypeError, ValueError):
-            pass  # unparseable ts: proceed, do not fail closed on metadata
-
-    # COMPLETED: >=2 distinct weighted pattern hits, no negation, not stale.
-    completed_hits = 0
-    completed_conf = 0.0
-    if completed_allowed and not _is_negated(text):
-        weight, distinct, _ = _completed_weight(text)
-        if weight >= COMPLETED_MIN_WEIGHT:
-            completed_hits = weight
-            # legacy confidence kept for API compat; NOT a reliability measure
-            completed_conf = min(0.95, 0.6 + (distinct - 1) * 0.2)
-
-    candidates = [
-        ("COMPLETED", completed_conf, completed_hits),
-        ("BLOCKER", _match_score(text, BLOCKER_PATTERNS),
-         sum(1 for p in BLOCKER_PATTERNS if p.search(text))),
-        ("DECISION", _match_score(text, DECISION_PATTERNS),
-         sum(1 for p in DECISION_PATTERNS if p.search(text))),
-        ("ALERT", _match_score(text, ALERT_PATTERNS),
-         sum(1 for p in ALERT_PATTERNS if p.search(text))),
-        ("MILESTONE", _match_score(text, MILESTONE_PATTERNS),
-         sum(1 for p in MILESTONE_PATTERNS if p.search(text))),
-    ]
-
-    # Sort by confidence descending; ALERT wins ties (safety: urgency first)
-    # Use negative confidence for descending, and ALERT as tiebreaker
-    candidates.sort(key=lambda x: (-x[1], 0 if x[0] == "ALERT" else 1))
-    best_cat, best_conf, best_hits = candidates[0]
-
-    if best_conf < min_confidence:
-        return None
-
-    return SiphonHit(
-        category=best_cat,
-        confidence=best_conf,
-        hits=best_hits,
-        summary=_extract_summary(text),
-        thread_id=thread_id,
-        message_id=message_id,
-    )
-
-
-def format_siphon(hit: SiphonHit, agent_name: str = "sidechat") -> str:
-    """
-    Format a siphon message for main chat.
-    HONEST LABELING: reports keyword hit count, never a fake "confidence %".
-    HONEST AUTHORSHIP (integrator, agent 3 absent): attributes the message's
-    real author when known. Falls back to the thread's registered agent only
-    when the author is unknown — and says so explicitly, so a relay can
-    never again launder thread ownership as authorship.
-    """
-    emoji = {"COMPLETED": "✅", "BLOCKER": "🚧", "DECISION": "❓",
-             "ALERT": "🚨", "MILESTONE": "🎯"}.get(hit.category, "📋")
-    thread_url = f"https://muse.ai/thread/{hit.thread_id}"
-    if hit.author:
-        attribution = f"from {hit.author}"
-    else:
-        attribution = f"from {agent_name} side chat (author unverified)"
-    return (
-        f"{emoji} [{hit.category}] {attribution}\n"
-        f"{hit.summary}\n"
-        f"→ {thread_url}\n"
-        f"(keyword hits: {hit.hits})"
-    )
-
-
-# --- Opt-out registry ---
-
-_opt_out_threads: set = set()
-
-
-def opt_out(thread_id: str):
-    """Agent opts a side chat out of siphoning."""
-    _opt_out_threads.add(thread_id)
-
-
-def opt_in(thread_id: str):
-    """Re-enable siphoning for a side chat."""
-    _opt_out_threads.discard(thread_id)
-
-
-def is_opted_out(thread_id: str) -> bool:
-    return thread_id in _opt_out_threads

bin/digest.py

@@ -1,74 +0,0 @@
-#!/usr/bin/env python3
-"""
-Side-chat work digest — INTEGRATOR minimal version (agent 4 of 5 never
-delivered its digest/throttle design within the window).
-
-Purpose: stop the per-message ✅ COMPLETED relay flood. COMPLETED hits are
-batched here and emitted as ONE periodic digest instead of N main-chat
-messages. ALERT / BLOCKER / DECISION still relay individually via
-siphon() — urgency is never batched.
-
-Interface (what agent 4's full design should remain compatible with):
-  - buffer = DigestBuffer(max_items=20, max_age_s=3600)
-  - buffer.add(hit) -> None
-  - buffer.flush() -> Optional[str]  (formatted digest, clears buffer)
-  - flush_digest() -> Optional[str]  (module-level singleton convenience)
-
-Reversible: to restore per-message COMPLETED relays, route COMPLETED back
-through siphon() in monitor.py and ignore this module.
-"""
-
-import time
-from typing import List, Optional
-
-try:
-    from detect import SiphonHit
-except ImportError:  # pragma: no cover
-    SiphonHit = object
-
-
-class DigestBuffer:
-    """Batch COMPLETED hits; flush() renders one digest message."""
-
-    def __init__(self, max_items: int = 20, max_age_s: int = 3600):
-        self.max_items = max_items
-        self.max_age_s = max_age_s
-        self._items: List[tuple] = []  # (ts, SiphonHit)
-
-    def add(self, hit) -> None:
-        now = time.time()
-        # Prune items older than max_age_s on every add (bounded memory).
-        self._items = [(ts, h) for ts, h in self._items
-                       if now - ts < self.max_age_s]
-        self._items.append((now, hit))
-        # Bound the buffer; oldest evicted first.
-        self._items = self._items[-self.max_items:]
-
-    def __len__(self) -> int:
-        return len(self._items)
-
-    def flush(self) -> Optional[str]:
-        """Render and clear. Returns None when there's nothing to digest."""
-        if not self._items:
-            return None
-        lines = ["📦 [DIGEST] completions from side chats "
-                 f"({len(self._items)} item(s))"]
-        for _, hit in self._items:
-            author = getattr(hit, "author", "") or "?"
-            url = f"https://muse.ai/thread/{hit.thread_id}"
-            lines.append(f"• {hit.summary} — {author} ({url})")
-        self._items = []
-        return "\n".join(lines)
-
-
-# Module-level singleton: the monitor loop shares one buffer per process.
-_default_buffer = DigestBuffer()
-
-
-def get_buffer() -> DigestBuffer:
-    return _default_buffer
-
-
-def flush_digest() -> Optional[str]:
-    """Flush the process-wide digest buffer. None if empty."""
-    return _default_buffer.flush()

bin/dm-log-taxonomy.py

@@ -1,267 +0,0 @@
-#!/usr/bin/env python3
-"""dm-log-taxonomy.py — READ-ONLY failure taxonomy for fleet DM sidechat reliability.
-
-Reads /home/super/Projects/NetVM/dm-log.jsonl, prints:
-  1. Event-type counts and send outcome rates
-  2. Sidechat nav failure taxonomy (per-target, per-agent-pair)
-  3. Failure timeline (hourly buckets, worst 10-min windows, by node)
-  4. "Ghost" rate: verified:true sidechat sends with no UUID anywhere
-  5. Hypothesis evidence tables (nav_failed reasons, placement pairs,
-     alias sources, autoprovision success, retry distribution)
-
-No writes to any state files. Runs in <1s on the current log size.
-"""
-import json
-import re
-import sys
-from collections import Counter, defaultdict
-from datetime import datetime, timezone, timedelta
-
-LOG = "/home/super/Projects/NetVM/dm-log.jsonl"
-WINDOW_H = 48
-
-
-def parse_ts(s):
-    if not s:
-        return None
-    try:
-        dt = datetime.fromisoformat(s.replace("Z", "+00:00"))
-        if dt.tzinfo is None:
-            dt = dt.replace(tzinfo=timezone.utc)
-        return dt
-    except Exception:
-        return None
-
-
-def main():
-    now = datetime.now(timezone.utc)
-    cutoff = now - timedelta(hours=WINDOW_H)
-    events = []
-    parse_err = 0
-    with open(LOG, encoding="utf-8") as f:
-        for line in f:
-            line = line.strip()
-            if not line:
-                continue
-            try:
-                e = json.loads(line)
-            except Exception:
-                parse_err += 1
-                continue
-            e["_dt"] = parse_ts(e.get("ts"))
-            events.append(e)
-    in_win = [e for e in events if e["_dt"] and e["_dt"] >= cutoff]
-    print(f"dm-log.jsonl: {len(events)} total lines ({parse_err} parse errors)")
-    print(f"window: last {WINDOW_H}h -> {len(in_win)} events")
-    if in_win:
-        print(f"  range: {in_win[0]['ts']} .. {in_win[-1]['ts']}")
-    print("=" * 78)
-
-    # ---- 1. event-type counts ------------------------------------------------
-    types = Counter(e.get("type", "?") for e in in_win)
-    print("\n[1] EVENT-TYPE COUNTS")
-    for t, n in types.most_common():
-        print(f"  {n:6d}  {t}")
-
-    # ---- per-send assembly ---------------------------------------------------
-    sends = {}
-    order = []
-
-    def S(e):
-        sid = e.get("id")
-        if not sid:
-            return None
-        if sid not in sends:
-            sends[sid] = {"id": sid, "events": [], "first_ts": e["_dt"]}
-            order.append(sid)
-        s = sends[sid]
-        s["events"].append(e)
-        for k in ("agent", "to", "target"):
-            if k not in s and e.get(k) is not None:
-                s[k] = e.get(k)
-        t = e.get("type")
-        if t == "verified":
-            s["verified_seen"] = True
-            if e.get("thread_uuid"):
-                s["uuids"] = s.get("uuids", set()) | {e["thread_uuid"]}
-        if t == "sent":
-            s["sent_seen"] = True
-            s["sent_verified"] = bool(e.get("verified"))
-            if e.get("thread_uuid"):
-                s["uuids"] = s.get("uuids", set()) | {e["thread_uuid"]}
-            tags = e.get("tags") or {}
-            if isinstance(tags, dict) and tags.get("thread"):
-                s["uuids"] = s.get("uuids", set()) | {tags["thread"]}
-        if t == "send_done":
-            s["done"] = True
-        if t == "sidechat_uuid_capture_failed":
-            s["capture_failed"] = True
-            s["capture_out"] = (e.get("out") or "")[:80]
-        if t == "sidechat_autoprovisioned" and e.get("thread_uuid"):
-            s["uuids"] = s.get("uuids", set()) | {e["thread_uuid"]}
-        if t == "alias_resolved" and e.get("thread_uuid"):
-            s["uuids"] = s.get("uuids", set()) | {e["thread_uuid"]}
-        if t == "placement_mismatch":
-            s["placement_mismatch"] = True
-        if t == "pre_send_assert_failed":
-            s["gate_failed"] = True
-            s["gate_reason"] = e.get("reason")
-        if t == "nav_failed":
-            s["nav_failed"] = True
-            s["nav_reason"] = e.get("reason") or "transport"
-        return s
-
-    for e in in_win:
-        S(e)
-
-    started = [sends[i] for i in order
-               if any(e.get("type") == "send_start" for e in sends[i]["events"])]
-    print(f"\n[1b] SEND OUTCOMES: {len(started)} sends attempted")
-    oc = Counter()
-    for s in started:
-        if s.get("gate_failed"):
-            oc["loud-fail: pre_send_assert_failed"] += 1
-        elif s.get("nav_failed"):
-            oc["loud-fail: nav_failed"] += 1
-        elif s.get("sent_seen") and s.get("sent_verified"):
-            oc["sent verified:true"] += 1
-        elif s.get("sent_seen"):
-            oc["sent verified:false"] += 1
-        elif s.get("done"):
-            oc["send_done, no sent event"] += 1
-        else:
-            oc["abandoned (no terminal event)"] += 1
-    for k, n in oc.most_common():
-        print(f"  {n:5d}  {k}")
-    print("  (note: main_chat_blocked events are policy blocks, not failures;")
-    print("   followup_register_failed=signing issues, tangential)")
-
-    # ---- 2. sidechat taxonomy ------------------------------------------------
-    def is_sc(s):
-        return (s.get("target") or "") not in ("main", "", None)
-
-    sc = [s for s in started if is_sc(s)]
-    print(f"\n[2] SIDECHAT TAXONOMY ({len(sc)} sidechat-targeted sends)")
-    tax = Counter()
-    per_target = defaultdict(Counter)
-    per_pair = defaultdict(Counter)
-    for s in sc:
-        if s.get("gate_failed"):
-            cls = "gate_failed:" + str(s.get("gate_reason"))
-        elif s.get("nav_failed"):
-            cls = "nav_failed:" + str(s.get("nav_reason"))
-        elif s.get("capture_failed"):
-            cls = "uuid_capture_failed"
-        elif s.get("placement_mismatch"):
-            cls = "placement_mismatch"
-        elif s.get("sent_seen") and s.get("sent_verified"):
-            cls = ("verified:true, NO uuid anywhere (ghost)"
-                   if not s.get("uuids") else "clean verified")
-        elif s.get("sent_seen"):
-            cls = "sent verified:false"
-        elif s.get("done"):
-            cls = "done, no sent event"
-        else:
-            cls = "abandoned"
-        tax[cls] += 1
-        per_target[s.get("target") or "?"][cls] += 1
-        per_pair[f"{s.get('agent') or '?'}->{s.get('to') or '?'}"][cls] += 1
-    for k, n in tax.most_common():
-        print(f"  {n:5d}  {k}")
-    print("\n  per-target (sends, non-clean, top classes):")
-    for tgt, c in sorted(per_target.items(), key=lambda x: -sum(x[1].values())):
-        tot = sum(c.values())
-        bad = tot - c.get("clean verified", 0)
-        print(f"    {tgt}: {tot} sends, {bad} non-clean {dict(c.most_common(4))}")
-    print("\n  per agent-pair:")
-    for pair, c in sorted(per_pair.items(), key=lambda x: -sum(x[1].values())):
-        tot = sum(c.values())
-        bad = tot - c.get("clean verified", 0)
-        print(f"    {pair}: {tot} sends, {bad} non-clean {dict(c.most_common(4))}")
-
-    # ---- 4. ghosts ------------------------------------------------------------
-    ghosts = [s for s in sc if s.get("sent_seen") and s.get("sent_verified")
-              and not s.get("uuids")]
-    print(f"\n[4] GHOST RATE: {len(ghosts)}/{len(sc)} "
-          f"({100.0 * len(ghosts) / len(sc) if sc else 0:.0f}%) verified:true "
-          f"sidechat sends with no UUID in any event")
-    gh = Counter(s["first_ts"].strftime("%m-%d %H") for s in ghosts if s["first_ts"])
-    print("  ghost hours:", dict(sorted(gh.items())))
-    print("  proxies: uuid_capture_failed="
-          f"{sum(1 for s in sc if s.get('capture_failed'))}, "
-          f"placement_mismatch={sum(1 for s in sc if s.get('placement_mismatch'))}, "
-          f"pre_send_assert_failed={sum(1 for s in sc if s.get('gate_failed'))}")
-
-    # ---- 3. timeline ------------------------------------------------------------
-    print("\n[3] TIMELINE (hourly, sidechat sends; #=bad, .=ok)")
-    buckets = defaultdict(Counter)
-    for s in sc:
-        if not s.get("first_ts"):
-            continue
-        hr = s["first_ts"].strftime("%m-%d %H:00")
-        buckets[hr]["total"] += 1
-        bad = not (s.get("sent_seen") and s.get("sent_verified")
-                   and not s.get("capture_failed")
-                   and not s.get("placement_mismatch")
-                   and not s.get("gate_failed") and not s.get("nav_failed"))
-        if bad:
-            buckets[hr]["bad"] += 1
-    for hr in sorted(buckets):
-        t, b = buckets[hr]["total"], buckets[hr]["bad"]
-        print(f"    {hr}  total={t:3d} bad={b:3d}  {'#' * b}{'.' * (t - b)}")
-    wins = defaultdict(Counter)
-    for s in sc:
-        if not s.get("first_ts"):
-            continue
-        w = s["first_ts"].strftime("%m-%d %H:%M")[:-1] + "0"
-        wins[w]["total"] += 1
-        if not (s.get("sent_seen") and s.get("sent_verified")):
-            wins[w]["bad"] += 1
-    print("  worst 10-min windows (>=3 sends):")
-    shown = 0
-    for w, c in sorted(wins.items(), key=lambda x: -x[1]["bad"]):
-        if c["total"] >= 3 and shown < 8:
-            print(f"    {w}  total={c['total']} bad={c['bad']}")
-            shown += 1
-    print("  bad rate by sending node:")
-    by_node = defaultdict(Counter)
-    for s in sc:
-        by_node[s.get("agent") or "?"]["total"] += 1
-        if not (s.get("sent_seen") and s.get("sent_verified")):
-            by_node[s.get("agent") or "?"]["bad"] += 1
-    for node, c in sorted(by_node.items(), key=lambda x: -x[1]["total"]):
-        r = 100.0 * c["bad"] / c["total"] if c["total"] else 0
-        print(f"    {node}: {c['bad']}/{c['total']} bad ({r:.0f}%)")
-
-    # ---- 5. hypothesis evidence ---------------------------------------------------
-    print("\n[5] HYPOTHESIS EVIDENCE")
-    nfr = Counter(s.get("nav_reason") for s in sc if s.get("nav_failed"))
-    print(f"  nav_failed reasons: {dict(nfr)}")
-    land = [s for s in sc if s.get("capture_failed")]
-    print(f"  uuid_capture_failed: {len(land)}, "
-          f"landing-page outs: {sum(1 for s in land if 'muse.ai/' in (s.get('capture_out') or ''))}")
-    pairs = Counter()
-    for e in in_win:
-        if e.get("type") == "placement_mismatch":
-            pairs[((e.get("expected_uuid") or "?")[:8],
-                   (e.get("actual_uuid") or "?")[:8], e.get("target"))] += 1
-    print("  placement_mismatch expected->actual:")
-    for (a, b, t), n in pairs.most_common(6):
-        print(f"    {n:3d}  exp={a}.. act={b}.. target={t}")
-    ar = Counter(e.get("source") for e in in_win if e.get("type") == "alias_resolved")
-    print(f"  alias_resolved sources: {dict(ar)} (None = field absent, older events)")
-    ap_try = sum(1 for e in in_win if e.get("type") == "sidechat_autoprovision_start")
-    ap_ok = sum(1 for e in in_win if e.get("type") == "sidechat_autoprovisioned"
-                and e.get("thread_uuid"))
-    print(f"  autoprovision: {ap_try} attempts -> {ap_ok} with uuid ({100.0 * ap_ok / ap_try if ap_try else 0:.0f}%)")
-    rt = Counter()
-    for s in started:
-        n = sum(1 for e in s["events"] if e.get("type") == "retry")
-        if n:
-            rt[n] += 1
-    print(f"  retry distribution (sends with >=1 retry): {dict(sorted(rt.items()))}")
-    print("\n  DONE.")
-
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/dm-sign.sh

@@ -1,74 +0,0 @@
-#!/usr/bin/env bash
-# dm-sign.sh — produce a signed DM for the fleet DM system.
-#
-# Signs a message with an SSH key (namespace "dm", Ed25519) and prints the
-# signed wire format that `dm.py verify-sig` checks:
-#
-#   [from:<identity>] [id:<id>]
-#
-#   <message body>
-#
-#   -----BEGIN SSH SIGNATURE-----
-#   ...
-#   -----END SSH SIGNATURE-----
-#
-# The signed payload is exactly "[from:X] [id:Y]\n\n<body>" (no trailing
-# newline) — verify-sig reconstructs it by stripping everything from the
-# signature block onward, so the two must match byte-for-byte.
-#
-# Usage:
-#   dm-sign.sh --from <identity> [--key <privkey>] [--id <id>] <message>
-#
-# Defaults: --key ~/.ssh/id_frontdoor, --id = 8 random hex chars.
-# The private key is only ever read locally; it is never moved or copied.
-# Pipe the output straight into `dm.py send --raw` (never truncate it).
-#
-# Example:
-#   dm.py send --agent opm --target main --raw \
-#     "$(dm-sign.sh --from operator-main 'hello from the operator')"
-set -euo pipefail
-
-FROM=""
-KEY="$HOME/.ssh/id_frontdoor"
-ID="$(head -c4 /dev/urandom | od -An -tx1 | tr -d ' \n')"
-
-usage() {
-    sed -n '2,/^set -euo/p' "$0" | sed 's/^# \?//'
-}
-
-while [[ $# -gt 0 ]]; do
-    case "$1" in
-        --from) FROM="${2:?--from needs a value}"; shift 2 ;;
-        --key)  KEY="${2:?--key needs a value}"; shift 2 ;;
-        --id)   ID="${2:?--id needs a value}"; shift 2 ;;
-        -h|--help) usage; exit 0 ;;
-        --) shift; break ;;
-        -*) echo "error: unknown option: $1" >&2; exit 1 ;;
-        *) break ;;
-    esac
-done
-
-if [[ $# -eq 0 ]]; then
-    echo "error: no message given" >&2
-    echo "usage: dm-sign.sh --from <identity> [--key <privkey>] [--id <id>] <message>" >&2
-    exit 1
-fi
-MESSAGE="$*"
-
-[[ -n "$FROM" ]] || { echo "error: --from <identity> is required" >&2; exit 1; }
-[[ -f "$KEY" ]]  || { echo "error: private key not found: $KEY" >&2; exit 1; }
-
-TD="$(mktemp -d)"
-trap 'rm -rf "$TD"' EXIT
-PAYLOAD="$TD/payload"
-
-# Payload: header, blank line, body — NO trailing newline (verify-sig strips).
-printf '[from:%s] [id:%s]\n\n%s' "$FROM" "$ID" "$MESSAGE" > "$PAYLOAD"
-
-# Never reuse a stale signature: a leftover .sig from an earlier run would
-# silently sign the wrong payload (burned 20 minutes on the board, 2026-10-03).
-rm -f "$PAYLOAD.sig"
-ssh-keygen -Y sign -f "$KEY" -n dm "$PAYLOAD" >/dev/null
-
-printf '[from:%s] [id:%s]\n\n%s\n\n' "$FROM" "$ID" "$MESSAGE"
-cat "$PAYLOAD.sig"

bin/docs-lookup

@@ -1 +0,0 @@
-/home/super/Projects/NetVM/bin/docs-lookup.py

bin/docs-lookup.py

@@ -1,606 +0,0 @@
-#!/usr/bin/env python3
-"""docs-lookup.py — Unified Lookup & Regex Passing Tool for docs_internal/.
-
-Provides high-speed queries, regex passing, grammar validation, and surface
-lookups for autonomous agents and operators interfacing with NetVM, Box,
-and box.muse-dev.online.
-
-Usage:
-  docs-lookup.py search <query>
-  docs-lookup.py surfaces [name]
-  docs-lookup.py sentence [name]
-  docs-lookup.py regex [name] [--test "<string>"]
-  docs-lookup.py parse "<string>"
-  docs-lookup.py get <collection> [key]
-  docs-lookup.py cli [domain]
-  docs-lookup.py overview
-"""
-
-import argparse
-import json
-import os
-import re
-import sys
-from pathlib import Path
-from typing import Any, Dict, List, Optional, Tuple
-
-NETVM_ROOT = Path(__file__).resolve().parent.parent
-LOOKUP_INTERNAL = NETVM_ROOT / "lookup_internal"
-if not LOOKUP_INTERNAL.exists() and (NETVM_ROOT / "docs_internal").exists():
-    LOOKUP_INTERNAL = NETVM_ROOT / "docs_internal"
-DOCS_INTERNAL = LOOKUP_INTERNAL
-
-USE_COLOR = sys.stdout.isatty() and os.environ.get("NO_COLOR") is None
-
-def _c(code: str, text: str) -> str:
-    return f"\033[{code}m{text}\033[0m" if USE_COLOR else str(text)
-
-def c_bold(s: str) -> str: return _c("1", s)
-def c_dim(s: str) -> str: return _c("2", s)
-def c_green(s: str) -> str: return _c("32", s)
-def c_red(s: str) -> str: return _c("31", s)
-def c_yellow(s: str) -> str: return _c("33", s)
-def c_blue(s: str) -> str: return _c("34", s)
-def c_cyan(s: str) -> str: return _c("36", s)
-def c_magenta(s: str) -> str: return _c("35", s)
-
-
-def load_json_file(filename: str) -> Dict[str, Any]:
-    """Safely load a JSON file from docs_internal."""
-    p = DOCS_INTERNAL / filename
-    if not p.is_file():
-        return {}
-    try:
-        with open(p, "r", encoding="utf-8") as f:
-            return json.load(f)
-    except Exception as e:
-        print(f"Error reading {p}: {e}", file=sys.stderr)
-        return {}
-
-
-def load_manifest() -> Dict[str, Any]:
-    return load_json_file("manifest.json")
-
-
-def load_sentence_structures() -> Dict[str, Any]:
-    return load_json_file("sentence_structure.json")
-
-
-def load_regex_patterns() -> Dict[str, Any]:
-    return load_json_file("regex_patterns.json")
-
-
-def load_assistive_surfaces() -> Dict[str, Any]:
-    return load_json_file("assistive_surfaces.json")
-
-
-def load_cli_tools() -> Dict[str, Any]:
-    return load_json_file("cli_tools.json")
-
-
-def load_fleet_nodes() -> Dict[str, Any]:
-    return load_json_file("fleet_nodes.json")
-
-
-def compile_pattern(pat_entry: Dict[str, Any]) -> Tuple[Optional[re.Pattern], Optional[str]]:
-    """Compile a regex entry with its configured flags."""
-    raw_pat = pat_entry.get("pattern", "")
-    flag_names = pat_entry.get("flags", [])
-    flags = 0
-    for fn in flag_names:
-        if hasattr(re, fn):
-            flags |= getattr(re, fn)
-    try:
-        return re.compile(raw_pat, flags), None
-    except Exception as e:
-        return None, str(e)
-
-
-# ---------------------------------------------------------------------------
-# Core Lookup Actions
-# ---------------------------------------------------------------------------
-
-def handle_overview(json_mode: bool = False):
-    manifest = load_manifest()
-    if json_mode:
-        print(json.dumps(manifest, indent=2))
-        return
-
-    print(c_bold("\n=== docs_internal — Internal Agent & Operator Lookup Database ==="))
-    print(c_dim(f"Location: {DOCS_INTERNAL} | Host: {manifest.get('host', 'box.muse-dev.online')}"))
-    print(f"{manifest.get('description', '')}\n")
-
-    print(c_cyan("Available Collections:"))
-    collections = manifest.get("collections", [])
-    for col in collections:
-        cid = col.get("id")
-        name = col.get("name")
-        desc = col.get("description")
-        jfile = col.get("json_file")
-        mfile = col.get("md_file")
-        print(f"  • {c_bold(cid):<20} {c_green(name)}")
-        print(f"    {c_dim(desc)}")
-        print(f"    {c_dim('Files:')} {c_yellow(jfile)} | {c_yellow(mfile)}")
-        print()
-
-    print(c_dim("Query commands: super docs [search|surfaces|sentence|regex|parse|get|cli]"))
-
-
-def handle_search(query: str, json_mode: bool = False):
-    q = query.lower()
-    results = []
-
-    # Search in all JSON files
-    for jpath in sorted(DOCS_INTERNAL.glob("*.json")):
-        try:
-            data = json.loads(jpath.read_text(encoding="utf-8"))
-        except Exception:
-            continue
-
-        def recurse_search(obj, path=""):
-            if isinstance(obj, dict):
-                for k, v in obj.items():
-                    subpath = f"{path}.{k}" if path else k
-                    if q in str(k).lower():
-                        results.append({
-                            "file": jpath.name,
-                            "type": "json_key",
-                            "path": subpath,
-                            "match": str(k),
-                            "preview": str(v)[:160]
-                        })
-                    recurse_search(v, subpath)
-            elif isinstance(obj, list):
-                for idx, item in enumerate(obj):
-                    recurse_search(item, f"{path}[{idx}]")
-            elif isinstance(obj, str):
-                if q in obj.lower():
-                    results.append({
-                        "file": jpath.name,
-                        "type": "json_value",
-                        "path": path,
-                        "match": obj[:120],
-                        "preview": obj[:240]
-                    })
-
-        recurse_search(data)
-
-    # Search in Markdown files
-    for mpath in sorted(DOCS_INTERNAL.glob("*.md")):
-        try:
-            content = mpath.read_text(encoding="utf-8")
-        except Exception:
-            continue
-        lines = content.splitlines()
-        for idx, line in enumerate(lines, 1):
-            if q in line.lower():
-                results.append({
-                    "file": mpath.name,
-                    "type": "markdown",
-                    "line": idx,
-                    "match": line.strip(),
-                    "preview": line.strip()
-                })
-
-    if json_mode:
-        print(json.dumps({"query": query, "count": len(results), "results": results}, indent=2))
-        return
-
-    print(c_bold(f"\nSearch results for '{query}' ({len(results)} matches):"))
-    if not results:
-        print(c_dim("  (no matching entries found)"))
-        return
-
-    for r in results[:40]:
-        if r["type"] == "markdown":
-            print(f"  [{c_yellow(r['file'])}:{c_cyan(str(r['line']))}] {r['match']}")
-        else:
-            print(f"  [{c_blue(r['file'])}:{c_magenta(r['path'])}] {r['preview']}")
-
-
-def handle_surfaces(view_name: Optional[str] = None, json_mode: bool = False):
-    data = load_assistive_surfaces()
-    views = data.get("views", {})
-
-    if view_name:
-        key = view_name.lower().strip()
-        v = views.get(key)
-        if not v:
-            for k, val in views.items():
-                if key in k or key in val.get("name", "").lower():
-                    v = val
-                    key = k
-                    break
-        if not v:
-            err = {"error": f"Surface '{view_name}' not found", "available": list(views.keys())}
-            if json_mode:
-                print(json.dumps(err, indent=2))
-            else:
-                print(c_red(f"Error: Surface '{view_name}' not found. Available: {', '.join(views.keys())}"))
-            sys.exit(1)
-
-        if json_mode:
-            print(json.dumps({key: v}, indent=2))
-            return
-
-        print(c_bold(f"\n=== Assistive Surface: {v.get('name')} (`{key}`) ==="))
-        print(c_dim(f"Host: {data.get('host')} | Tab ID: {v.get('tab_id')}"))
-        print(f"{c_cyan('DOM Tab Selector:')}  {c_yellow(str(v.get('dom_tab_selector')))}")
-        print(f"{c_cyan('DOM Pane Selector:')} {c_yellow(str(v.get('dom_pane_selector')))}")
-        
-        elements = v.get("key_elements", {})
-        if elements:
-            print(c_bold("\nKey DOM Elements / Selectors:"))
-            for el_name, sel in elements.items():
-                print(f"  • {c_magenta(el_name):<20} {c_green(sel)}")
-
-        endpoints = v.get("api_endpoints", [])
-        if endpoints:
-            print(c_bold("\nAssociated REST API Endpoints:"))
-            for ep in endpoints:
-                print(f"  • {c_bold(ep.get('method'))} {c_cyan(ep.get('path'))}")
-                print(f"    {c_dim(ep.get('description'))}")
-                if ep.get("curl_example"):
-                    print(f"    {c_dim('curl:')} {c_yellow(ep.get('curl_example'))}")
-
-        recipe = v.get("assistive_recipe")
-        if recipe:
-            print(c_bold("\nAssistive Recipe:"))
-            print(f"  {recipe}")
-        print()
-        return
-
-    # All views
-    if json_mode:
-        print(json.dumps(data, indent=2))
-        return
-
-    print(c_bold(f"\n=== Assistive Surfaces for {data.get('host', 'box.muse-dev.online')} ==="))
-    print(c_dim(f"Operator PIN: {data.get('auth', {}).get('pin')} | Session Cookie: {data.get('auth', {}).get('cookie_name')}"))
-    print()
-
-    for k, v in views.items():
-        name = v.get("name", k)
-        tab_sel = v.get("dom_tab_selector") or "(modal/overlay)"
-        eps = [f"{ep.get('method')} {ep.get('path')}" for ep in v.get("api_endpoints", [])]
-        ep_summary = ", ".join(eps) if eps else "(no direct endpoint)"
-        print(f"  • {c_bold(k):<16} {c_cyan(name):<30} {c_yellow(tab_sel)}")
-        print(f"    {c_dim('API:')} {ep_summary}")
-        if v.get("assistive_recipe"):
-            print(f"    {c_dim('Hint:')} {v.get('assistive_recipe')[:100]}...")
-        print()
-
-
-def handle_sentence(name: Optional[str] = None, json_mode: bool = False):
-    data = load_sentence_structures()
-    structs = data.get("structures", {})
-
-    if name:
-        key = name.lower().strip()
-        s = structs.get(key)
-        if not s:
-            for k, val in structs.items():
-                if key in k or key in val.get("name", "").lower() or key in val.get("protocol_tag", "").lower():
-                    s = val
-                    key = k
-                    break
-        if not s:
-            err = {"error": f"Sentence structure '{name}' not found", "available": list(structs.keys())}
-            if json_mode:
-                print(json.dumps(err, indent=2))
-            else:
-                print(c_red(f"Error: Sentence structure '{name}' not found. Available: {', '.join(structs.keys())}"))
-            sys.exit(1)
-
-        if json_mode:
-            print(json.dumps({key: s}, indent=2))
-            return
-
-        print(c_bold(f"\n=== Protocol Structure: {s.get('name')} (`{key}`) ==="))
-        print(f"{c_cyan('Tag:')}        {c_bold(s.get('protocol_tag'))}")
-        print(f"{c_cyan('Template:')}   {c_green(s.get('template'))}")
-        print(f"{c_cyan('Lifecycle:')}  {c_yellow(s.get('lifecycle_transition', ''))}")
-        print(f"\n{c_bold('Description:')}\n  {s.get('description')}")
-        print(f"\n{c_bold('Required Fields:')} {', '.join(s.get('required_fields', []))}")
-        if s.get("optional_fields"):
-            print(f"{c_bold('Optional Fields:')} {', '.join(s.get('optional_fields', []))}")
-        print(f"\n{c_bold('Example:')}\n  {c_cyan(s.get('example'))}")
-        print(f"\n{c_bold('Reply Expectation:')}\n  {s.get('reply_expectation')}")
-        print()
-        return
-
-    if json_mode:
-        print(json.dumps(data, indent=2))
-        return
-
-    print(c_bold("\n=== Agent Sentence Structures & Conversational Contracts ==="))
-    print(c_dim("Format rules enforced by response-harvester and self_main_loop.\n"))
-
-    for k, s in structs.items():
-        tag = s.get("protocol_tag", "")
-        desc = s.get("description", "")
-        print(f"  • {c_bold(k):<18} {c_green(tag):<25} {s.get('name')}")
-        print(f"    {c_dim(desc)}")
-        print(f"    {c_dim('Template:')} {c_cyan(s.get('template', ''))}")
-        print()
-
-
-def handle_regex(name: Optional[str] = None, test_str: Optional[str] = None, json_mode: bool = False):
-    data = load_regex_patterns()
-    pats = data.get("patterns", {})
-
-    if name:
-        key = name.lower().strip()
-        p = pats.get(key)
-        if not p:
-            for k, val in pats.items():
-                if key in k or key in val.get("name", "").lower():
-                    p = val
-                    key = k
-                    break
-        if not p:
-            err = {"error": f"Regex pattern '{name}' not found", "available": list(pats.keys())}
-            if json_mode:
-                print(json.dumps(err, indent=2))
-            else:
-                print(c_red(f"Error: Pattern '{name}' not found. Available: {', '.join(pats.keys())}"))
-            sys.exit(1)
-
-        compiled, comp_err = compile_pattern(p)
-
-        test_result = None
-        if test_str is not None:
-            if compiled:
-                m = compiled.search(test_str)
-                test_result = {
-                    "matched": bool(m),
-                    "match_span": m.span() if m else None,
-                    "matched_text": m.group(0) if m else None,
-                    "named_groups": m.groupdict() if m else {},
-                    "groups": list(m.groups()) if m else []
-                }
-            else:
-                test_result = {"matched": False, "error": comp_err}
-
-        if json_mode:
-            out = {key: p, "compiled_ok": bool(compiled)}
-            if test_str is not None:
-                out["test_evaluation"] = test_result
-            print(json.dumps(out, indent=2))
-            return
-
-        print(c_bold(f"\n=== Regex Pattern: {p.get('name')} (`{key}`) ==="))
-        print(f"{c_cyan('Pattern:')}      {c_yellow(p.get('pattern'))}")
-        print(f"{c_cyan('Flags:')}        {', '.join(p.get('flags', [])) or '(none)'}")
-        print(f"{c_cyan('Description:')}  {p.get('description')}")
-        print(f"{c_cyan('Usage:')}        {c_dim(p.get('usage', ''))}")
-
-        ng = p.get("named_groups", {})
-        if ng:
-            print(c_bold("\nNamed Groups:"))
-            for gname, gdesc in ng.items():
-                print(f"  • {c_magenta(gname):<16} {gdesc}")
-
-        if test_str is not None:
-            print(c_bold("\nTest Execution Result:"))
-            print(f"  Input: {c_dim(test_str)}")
-            if test_result.get("matched"):
-                print(f"  Verdict: {c_green('✔ MATCHED')}")
-                print(f"  Matched Text: {c_cyan(test_result['matched_text'])}")
-                if test_result["named_groups"]:
-                    print(f"  Extracted Tokens:")
-                    for k_grp, v_grp in test_result["named_groups"].items():
-                        print(f"    - {c_magenta(k_grp)}: {c_yellow(str(v_grp))}")
-            else:
-                print(f"  Verdict: {c_red('✖ NO MATCH')}")
-                if comp_err:
-                    print(f"  Compile Error: {comp_err}")
-        print()
-        return
-
-    # List patterns
-    if json_mode:
-        print(json.dumps(data, indent=2))
-        return
-
-    print(c_bold("\n=== Master Regex Patterns & Passing Dictionary ==="))
-    print(c_dim("Patterns tested and calibrated across response-harvester, dm, and loop engines.\n"))
-
-    for k, p in pats.items():
-        print(f"  • {c_bold(k):<18} {c_green(p.get('name'))}")
-        print(f"    {c_dim('Pattern:')} {c_yellow(p.get('pattern'))}")
-        print(f"    {c_dim(p.get('description'))}")
-        print()
-
-
-def handle_parse(candidate_str: str, json_mode: bool = False):
-    """Pass candidate_str through all registered regexes and extract tokens."""
-    data = load_regex_patterns()
-    pats = data.get("patterns", {})
-
-    matches = []
-    for k, p in pats.items():
-        compiled, err = compile_pattern(p)
-        if not compiled:
-            continue
-        m = compiled.search(candidate_str)
-        if m:
-            matches.append({
-                "pattern_key": k,
-                "pattern_name": p.get("name"),
-                "matched_text": m.group(0),
-                "named_groups": m.groupdict(),
-                "span": m.span()
-            })
-
-    if json_mode:
-        print(json.dumps({
-            "input": candidate_str,
-            "matched_patterns_count": len(matches),
-            "matches": matches
-        }, indent=2))
-        return
-
-    print(c_bold(f"\n=== Regex Parse Analysis ==="))
-    print(f"Input: {c_dim(candidate_str)}\n")
-
-    if not matches:
-        print(c_yellow("  ⚠ No registered regex pattern matched this string."))
-        return
-
-    print(c_green(f"Matched {len(matches)} pattern(s):"))
-    for match in matches:
-        print(f"\n  • Pattern: {c_bold(match['pattern_name'])} (`{c_cyan(match['pattern_key'])}`)")
-        print(f"    Matched Chunk: {c_yellow(match['matched_text'])}")
-        ng = match["named_groups"]
-        if ng:
-            print(f"    Extracted Tokens:")
-            for gk, gv in ng.items():
-                print(f"      - {c_magenta(gk)}: {c_green(str(gv))}")
-    print()
-
-
-def handle_get(collection: str, key: Optional[str] = None, json_mode: bool = False):
-    col_map = {
-        "manifest": "manifest.json",
-        "sentence": "sentence_structure.json",
-        "sentence_structure": "sentence_structure.json",
-        "regex": "regex_patterns.json",
-        "regex_patterns": "regex_patterns.json",
-        "surfaces": "assistive_surfaces.json",
-        "assistive_surfaces": "assistive_surfaces.json",
-        "cli": "cli_tools.json",
-        "cli_tools": "cli_tools.json",
-        "fleet": "fleet_nodes.json",
-        "fleet_nodes": "fleet_nodes.json",
-    }
-
-    fname = col_map.get(collection.lower().strip())
-    if not fname:
-        print(c_red(f"Error: Unknown collection '{collection}'. Available: {', '.join(col_map.keys())}"), file=sys.stderr)
-        sys.exit(1)
-
-    data = load_json_file(fname)
-    if key:
-        # Check top level or primary container
-        val = None
-        for primary in ["structures", "patterns", "views", "domains", "nodes", "collections"]:
-            if primary in data and isinstance(data[primary], dict) and key in data[primary]:
-                val = data[primary][key]
-                break
-        if val is None and key in data:
-            val = data[key]
-        if val is None:
-            print(c_red(f"Error: Key '{key}' not found in {fname}"), file=sys.stderr)
-            sys.exit(1)
-        data = {key: val}
-
-    print(json.dumps(data, indent=2))
-
-
-def handle_cli(domain: Optional[str] = None, json_mode: bool = False):
-    data = load_cli_tools()
-    domains = data.get("domains", {})
-
-    if domain:
-        d = domains.get(domain.lower().strip())
-        if not d:
-            print(c_red(f"Error: CLI domain '{domain}' not found. Available: {', '.join(domains.keys())}"), file=sys.stderr)
-            sys.exit(1)
-        if json_mode:
-            print(json.dumps({domain: d}, indent=2))
-            return
-        print(c_bold(f"\n=== CLI Tool Domain: super {domain} / box {domain} ==="))
-        print(f"Summary: {d.get('summary')}\n")
-        for sub in d.get("subcommands", []):
-            print(f"  • {c_green(sub['cmd'])}")
-            print(f"    {c_dim(sub['desc'])}\n")
-        return
-
-    if json_mode:
-        print(json.dumps(data, indent=2))
-        return
-
-    print(c_bold("\n=== Unified NetVM & Box CLI Tool Catalog ==="))
-    print(c_dim("Powered by super-cli.py and bin/muse wrapper.\n"))
-    for dom_k, dom_v in domains.items():
-        print(f"  • {c_bold(dom_k):<16} {c_cyan(dom_v.get('summary'))}")
-        for sub in dom_v.get("subcommands", [])[:2]:
-            print(f"    - {c_dim(sub['cmd'])}")
-        print()
-
-
-# ---------------------------------------------------------------------------
-# CLI Argument Parser
-# ---------------------------------------------------------------------------
-
-def build_parser():
-    common = argparse.ArgumentParser(add_help=False)
-    common.add_argument("--json", action="store_true", help="Output machine-readable JSON")
-
-    parser = argparse.ArgumentParser(
-        description="docs-lookup — Internal Agent & Operator Documentation Query Engine",
-        parents=[common],
-        formatter_class=argparse.RawDescriptionHelpFormatter
-    )
-
-    sub = parser.add_subparsers(dest="action")
-
-    p_search = sub.add_parser("search", parents=[common], help="Full-text search across docs_internal")
-    p_search.add_argument("query", help="Search keyword or phrase")
-
-    p_surfaces = sub.add_parser("surfaces", parents=[common], help="Lookup assistive surfaces for box.muse-dev.online")
-    p_surfaces.add_argument("name", nargs="?", default=None, help="Surface or tab name")
-
-    p_sentence = sub.add_parser("sentence", parents=[common], help="Lookup agent sentence structures & protocols")
-    p_sentence.add_argument("name", nargs="?", default=None, help="Structure name (e.g. work_order, result)")
-
-    p_regex = sub.add_parser("regex", parents=[common], help="Lookup and test regex patterns")
-    p_regex.add_argument("name", nargs="?", default=None, help="Pattern name (e.g. work_order, verb)")
-    p_regex.add_argument("--test", dest="test_str", default=None, help="Test string to evaluate against pattern")
-
-    p_parse = sub.add_parser("parse", parents=[common], help="Parse an agent utterance through all regex patterns")
-    p_parse.add_argument("string", help="String/utterance to parse")
-
-    p_get = sub.add_parser("get", parents=[common], help="Query raw JSON collection and key")
-    p_get.add_argument("collection", help="Collection name (sentence, regex, surfaces, cli, fleet)")
-    p_get.add_argument("key", nargs="?", default=None, help="Optional specific key")
-
-    p_cli = sub.add_parser("cli", parents=[common], help="Lookup CLI commands and syntax")
-    p_cli.add_argument("domain", nargs="?", default=None, help="CLI domain (fleet, dm, job, etc.)")
-
-    p_overview = sub.add_parser("overview", parents=[common], help="Database overview and collections manifest")
-
-    return parser
-
-
-
-def main():
-    parser = build_parser()
-    args = parser.parse_args()
-
-    act = args.action
-    json_mode = getattr(args, "json", False)
-
-    if not act or act == "overview":
-        handle_overview(json_mode)
-    elif act == "search":
-        handle_search(args.query, json_mode)
-    elif act == "surfaces":
-        handle_surfaces(args.name, json_mode)
-    elif act == "sentence":
-        handle_sentence(args.name, json_mode)
-    elif act == "regex":
-        handle_regex(args.name, args.test_str, json_mode)
-    elif act == "parse":
-        handle_parse(args.string, json_mode)
-    elif act == "get":
-        handle_get(args.collection, args.key, json_mode)
-    elif act == "cli":
-        handle_cli(args.domain, json_mode)
-    else:
-        parser.print_help()
-
-
-if __name__ == "__main__":
-    main()

bin/ensure-node-supervision.sh

@@ -1,95 +0,0 @@
-#!/usr/bin/env bash
-# ensure-node-supervision.sh <node> | --all — feed a node to the watchdogs.
-#
-# Setup (netvm-node-up.sh, hence netvm-provision-node.sh and the onboarding
-# pipeline) calls this so every node gets supervision without manual wiring:
-#   1. NODES.md registry row (idempotent) — feeds the registry-driven
-#      supervisors: cdp-relay-watchdog, agent-health.sh, relay-health-check,
-#      cdp-latency-check. Port from netvm-names pinning (honors
-#      CDP_PORT_OVERRIDE, so provision's picked port wins when present).
-#   2. chromebox-watchdog-<node>.timer unit + enable --now — the one
-#      supervisor that needs a per-node systemd unit (the @.service
-#      template already exists). Needs root for the real unit dir.
-#
-# Env overrides (tests): NODES_MD, UNIT_DIR. systemctl is skipped when
-# UNIT_DIR is not the real system dir.
-#
-# Runs at the end of netvm-node-up.sh (as root); safe to re-run anytime:
-#   sudo bin/ensure-node-supervision.sh --all
-set -euo pipefail
-SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
-NODES_MD="${NODES_MD:-$SCRIPT_DIR/../NODES.md}"
-UNIT_DIR="${UNIT_DIR:-/etc/systemd/system}"
-# shellcheck disable=SC1091
-. "$SCRIPT_DIR/netvm-names.sh"
-
-usage() { echo "usage: ensure-node-supervision.sh <node> | --all" >&2; exit 1; }
-
-ensure_registry_row() {
-  local node="$1"
-  if grep -qE "^\|[[:space:]]*$node[[:space:]]*\|" "$NODES_MD" 2>/dev/null; then
-    echo "registry: $node already in NODES.md"
-    return 0
-  fi
-  netvm_names "$node" || { echo "registry: unknown node $node" >&2; return 1; }
-  printf '| %s | %s | unknown | %s | active | %s (auto-registered) |\n' \
-    "$node" "$NETNS" "$CDP_PORT" "$node" >> "$NODES_MD"
-  echo "registry: added $node (port $CDP_PORT)"
-}
-
-ensure_timer() {
-  local node="$1" unit
-  unit="$UNIT_DIR/chromebox-watchdog-$node.timer"
-  if [ -f "$unit" ]; then
-    echo "timer: chromebox-watchdog-$node.timer already installed"
-  else
-    if [ "$UNIT_DIR" = "/etc/systemd/system" ] && [ "$(id -u)" -ne 0 ]; then
-      echo "timer: need root to install chromebox-watchdog-$node.timer (run with sudo)" >&2
-      return 1
-    fi
-    cat > "$unit" <<EOF
-[Unit]
-Description=Run chromebox watchdog for $node every 2 minutes
-
-[Timer]
-RandomizedDelaySec=30s
-OnBootSec=2min
-OnUnitActiveSec=2min
-Unit=chromebox-watchdog@$node.service
-
-[Install]
-WantedBy=timers.target
-EOF
-    echo "timer: installed chromebox-watchdog-$node.timer"
-  fi
-  if [ "$UNIT_DIR" = "/etc/systemd/system" ]; then
-    systemctl daemon-reload
-    systemctl enable --now "chromebox-watchdog-$node.timer" >/dev/null 2>&1
-    echo "timer: enabled chromebox-watchdog-$node.timer"
-  fi
-}
-
-ensure_node() {
-  local node="$1"
-  ensure_registry_row "$node"
-  ensure_timer "$node"
-}
-
-case "${1:-}" in
-  --all)
-    nodes="$(python3 "$SCRIPT_DIR/netvm-registry.py" 2>/dev/null | cut -d: -f1)"
-    for conf in /etc/netvm/*.conf; do
-      [ -f "$conf" ] || continue
-      nodes="$nodes $(basename "$conf" .conf)"
-    done
-    seen=""
-    # shellcheck disable=SC2086 (intended word splitting)
-    for node in $nodes; do
-      case " $seen " in *" $node "*) continue;; esac
-      seen="$seen $node"
-      ensure_node "$node" || echo "supervision: $node failed (continuing)" >&2
-    done
-    ;;
-  ""|-h|--help) usage;;
-  *) ensure_node "$1";;
-esac

bin/exec-server.py

@@ -1,377 +0,0 @@
-#!/usr/bin/env python3
-"""
-HTTPS exec server for operator remote command execution.
-Runs on bl (stable), accepts authenticated POST /exec, returns command output.
-
-Usage:
-    python3 exec-server.py --port 8443 --token-file /home/super/.exec-token
-
-646 (or any operator) can then:
-    curl -k -X POST https://100.123.153.75:8443/exec \
-        -H "Content-Type: application/json" \
-        -d '{"cmd": "dm.py send --agent 646 --to opm --target main \"hello\"", "token": "..."}'
-
-Via the VM Caddy (no tailnet needed from the agent's container):
-    curl -k -X POST https://34-139-37-135.sslip.io/exec/exec \
-        -H "Content-Type: application/json" \
-        -d '{"cmd": "...", "token": "<per-agent token>"}'
-
-Signature auth (preferred for agents): no secret crosses the wire at all.
-The agent signs {"cmd","ts","nonce"} with their registered SSH key
-(`ssh-keygen -Y sign -n exec-server`) and posts
-{"identity","payload","signature"}. Verified against the signers file
-with `ssh-keygen -Y verify`; ts must be within 300s and the nonce unused.
-This is the same identity primitive as signed board posts, and it never
-trips secret-handling guardrails because a signature is not a secret.
-
-Auth: the master token (TOKEN_FILE) plus per-agent tokens, one file per
-agent under TOKEN_DIR (0600). Each agent's token is individually revocable
-by deleting its file. The using identity is logged (never the token).
-"""
-import argparse
-import hashlib
-import hmac
-import json
-import secrets
-import subprocess
-import sys
-from http.server import HTTPServer, BaseHTTPRequestHandler
-import ssl
-
-# Token file - generated on first run if not exists
-TOKEN_FILE = '/home/super/.exec-server-token'
-# Per-agent tokens: TOKEN_DIR/<agent> contains that agent's token
-TOKEN_DIR = '/home/super/.exec-tokens'
-# ssh-keygen signature auth: public keys of fleet identities, one per line
-# ("<identity> ssh-ed25519 AAAA..."), synced from the VM's allowed_signers.
-SIGNERS_FILE = '/home/super/.exec-signers'
-# Seen nonces for replay protection ("<nonce> <ts>" per line).
-NONCE_FILE = '/home/super/.exec-nonces'
-SIG_NAMESPACE = 'exec-server'
-SIG_MAX_SKEW = 300  # seconds; nonces remembered for 2x this
-
-def get_token():
-    """Load or generate the auth token."""
-    try:
-        with open(TOKEN_FILE, 'r') as f:
-            return f.read().strip()
-    except FileNotFoundError:
-        token = secrets.token_hex(32)
-        with open(TOKEN_FILE, 'w') as f:
-            f.write(token)
-        # Secure permissions
-        import os
-        os.chmod(TOKEN_FILE, 0o600)
-        print(f'Generated new token in {TOKEN_FILE}', file=sys.stderr)
-        return token
-
-def check_token(token):
-    """Check token against the master token and per-agent tokens.
-    Returns the identity label ('master' or the agent filename), or None."""
-    if token and hmac.compare_digest(token, get_token()):
-        return 'master'
-    import os
-    try:
-        names = os.listdir(TOKEN_DIR)
-    except FileNotFoundError:
-        return None
-    for name in names:
-        p = os.path.join(TOKEN_DIR, name)
-        if not os.path.isfile(p):
-            continue
-        try:
-            with open(p) as f:
-                t = f.read().strip()
-        except OSError:
-            continue
-        if t and hmac.compare_digest(token, t):
-            return name
-    return None
-
-def check_nonce(nonce):
-    """True if the nonce was never used; records it. Prunes expired entries."""
-    import os
-    import time
-    now = time.time()
-    fresh = []
-    try:
-        with open(NONCE_FILE) as f:
-            for line in f:
-                parts = line.split()
-                if len(parts) != 2:
-                    continue
-                n, t = parts
-                try:
-                    if now - float(t) < 2 * SIG_MAX_SKEW:
-                        fresh.append((n, t))
-                except ValueError:
-                    pass
-    except FileNotFoundError:
-        pass
-    if any(n == nonce for n, _ in fresh):
-        return False
-    fresh.append((nonce, str(now)))
-    try:
-        with open(NONCE_FILE, 'w') as f:
-            for n, t in fresh:
-                f.write(f"{n} {t}\n")
-        os.chmod(NONCE_FILE, 0o600)
-    except OSError:
-        return False
-    return True
-
-
-def check_signature(identity, payload, signature):
-    """Verify an `ssh-keygen -Y` signature over the payload envelope.
-    Returns the identity on success, None on failure. The envelope must be
-    JSON {"cmd","ts","nonce"} with a fresh ts and an unused nonce."""
-    import json as _json
-    import os
-    import re
-    import subprocess
-    import tempfile
-    import time
-    if not re.fullmatch(r'[a-z0-9-]+', identity or ''):
-        return None
-    try:
-        data = _json.loads(payload)
-    except Exception:
-        return None
-    if not isinstance(data, dict):
-        return None
-    cmd = data.get('cmd')
-    ts = data.get('ts')
-    nonce = data.get('nonce')
-    if not isinstance(cmd, str) or not cmd:
-        return None
-    if not isinstance(nonce, str) or not re.fullmatch(r'[0-9a-fA-F]{16,128}', nonce):
-        return None
-    try:
-        ts = float(ts)
-    except (TypeError, ValueError):
-        return None
-    if abs(time.time() - ts) > SIG_MAX_SKEW:
-        return None
-    if not check_nonce(nonce):
-        return None
-    sig_path = None
-    try:
-        with tempfile.NamedTemporaryFile('w', delete=False, suffix='.sig') as f:
-            f.write(signature if signature.endswith('\n') else signature + '\n')
-            sig_path = f.name
-        p = subprocess.run(
-            ['ssh-keygen', '-Y', 'verify', '-f', SIGNERS_FILE, '-I', identity,
-             '-n', SIG_NAMESPACE, '-s', sig_path],
-            input=payload.encode(), capture_output=True, timeout=15)
-        return identity if p.returncode == 0 else None
-    except Exception:
-        return None
-    finally:
-        if sig_path:
-            try:
-                os.unlink(sig_path)
-            except OSError:
-                pass
-
-
-class ExecHandler(BaseHTTPRequestHandler):
-    def log_message(self, format, *args):
-        # Quiet logging, just to stderr
-        sys.stderr.write(f'{self.client_address[0]} - {format % args}\n')
-    
-    def do_POST(self):
-        # Read body
-        content_length = int(self.headers.get('Content-Length', 0))
-        if content_length > 1024 * 1024:  # 1MB max
-            self.send_response(413)
-            self.end_headers()
-            return
-
-        body = self.rfile.read(content_length)
-        try:
-            data = json.loads(body)
-        except json.JSONDecodeError:
-            self.send_response(400)
-            self.end_headers()
-            self.wfile.write(b'{"error": "invalid json"}')
-            return
-
-        if self.path == '/exec/rotate':
-            self.handle_rotate(data)
-            return
-
-        if self.path != '/exec':
-            self.send_response(404)
-            self.end_headers()
-            return
-        
-        # Auth: bearer token OR ssh-keygen -Y signature (no secret in transit).
-        # Bearer: {"token": "...", "cmd": "..."}.
-        # Signature: {"identity": "...", "payload": "{\"cmd\":...,\"ts\":...,\"nonce\":...}",
-        #             "signature": "<ssh-keygen -Y armor>"}. check_signature
-        # validates the envelope; cmd comes from the signed payload.
-        ident = None
-        cmd = ''
-        token = data.get('token', '')
-        if token:
-            ident = check_token(token)
-            cmd = data.get('cmd', '')
-        elif data.get('identity') and data.get('payload') and data.get('signature'):
-            ident = check_signature(data['identity'], data['payload'], data['signature'])
-            if ident:
-                try:
-                    cmd = json.loads(data['payload']).get('cmd', '')
-                except Exception:
-                    cmd = ''
-        if not ident:
-            self.send_response(401)
-            self.end_headers()
-            self.wfile.write(b'{"error": "unauthorized"}')
-            return
-        sys.stderr.write(f'exec as {ident} from {self.client_address[0]}\n')
-        # Get command
-        if not cmd or not isinstance(cmd, str):
-            self.send_response(400)
-            self.end_headers()
-            self.wfile.write(b'{"error": "missing cmd"}')
-            return
-        
-        # Execute (with timeout)
-        timeout = min(data.get('timeout', 60), 300)  # max 5 min
-        try:
-            result = subprocess.run(
-                cmd,
-                shell=True,
-                capture_output=True,
-                text=True,
-                timeout=timeout,
-                cwd='/home/super'
-            )
-            response = {
-                'stdout': result.stdout,
-                'stderr': result.stderr,
-                'rc': result.returncode,
-            }
-        except subprocess.TimeoutExpired:
-            response = {'error': 'timeout', 'rc': -1}
-        except Exception as e:
-            response = {'error': str(e), 'rc': -1}
-        
-        # Send response
-        resp_body = json.dumps(response).encode()
-        self.send_response(200)
-        self.send_header('Content-Type', 'application/json')
-        self.send_header('Content-Length', str(len(resp_body)))
-        self.end_headers()
-        self.wfile.write(resp_body)
-    
-    def handle_rotate(self, data):
-        """POST /exec/rotate - replace an agent's token with a fresh one.
-        The old token dies immediately; the new one is returned ONLY in the
-        response body (never logged). Lets an agent bootstrap from a
-        trust-root-delivered token and end up with one nobody else knows.
-        Agents may rotate only their own token; master may rotate any
-        agent's (not its own - that stays manual on bl)."""
-        import os
-        import re
-        token = data.get('token', '')
-        ident = check_token(token)
-        if not ident:
-            self.send_response(401)
-            self.end_headers()
-            self.wfile.write(b'{"error": "unauthorized"}')
-            return
-        target = data.get('agent', ident)
-        if not re.fullmatch(r'[a-z0-9-]+', target or ''):
-            self.send_response(400)
-            self.end_headers()
-            self.wfile.write(b'{"error": "bad agent name"}')
-            return
-        if target == 'master' or (ident != 'master' and target != ident):
-            self.send_response(403)
-            self.end_headers()
-            self.wfile.write(b'{"error": "forbidden"}')
-            return
-        path = os.path.join(TOKEN_DIR, target)
-        if not os.path.isfile(path):
-            self.send_response(404)
-            self.end_headers()
-            self.wfile.write(b'{"error": "no such agent token"}')
-            return
-        new_token = secrets.token_hex(32)
-        tmp = path + '.tmp'
-        with open(tmp, 'w') as f:
-            f.write(new_token + '\n')
-        os.chmod(tmp, 0o600)
-        os.replace(tmp, path)
-        sys.stderr.write(f'token rotated for {target} by {ident}\n')
-        resp_body = json.dumps({'token': new_token}).encode()
-        self.send_response(200)
-        self.send_header('Content-Type', 'application/json')
-        self.send_header('Content-Length', str(len(resp_body)))
-        self.end_headers()
-        self.wfile.write(resp_body)
-
-    def do_GET(self):
-        if self.path == '/health':
-            self.send_response(200)
-            self.send_header('Content-Type', 'application/json')
-            self.end_headers()
-            self.wfile.write(b'{"status": "ok"}')
-        else:
-            self.send_response(404)
-            self.end_headers()
-
-def main():
-    global TOKEN_FILE, TOKEN_DIR
-    parser = argparse.ArgumentParser()
-    parser.add_argument('--port', type=int, default=8443)
-    parser.add_argument('--host', default='0.0.0.0')
-    parser.add_argument('--token-file', default='/home/super/.exec-server-token')
-    parser.add_argument('--token-dir', default='/home/super/.exec-tokens')
-    parser.add_argument('--signers-file', default='/home/super/.exec-signers')
-    parser.add_argument('--nonce-file', default='/home/super/.exec-nonces')
-    args = parser.parse_args()
-
-    TOKEN_FILE = args.token_file
-    TOKEN_DIR = args.token_dir
-    global SIGNERS_FILE, NONCE_FILE
-    SIGNERS_FILE = args.signers_file
-    NONCE_FILE = args.nonce_file
-    
-    # Ensure token exists
-    token = get_token()
-    print(f'Token: {token[:8]}... (full in {TOKEN_FILE})', file=sys.stderr)
-    
-    server = HTTPServer((args.host, args.port), ExecHandler)
-    
-    # Wrap with TLS (self-signed is fine for our use, we use -k)
-    # Generate self-signed cert if not exists
-    import os
-    cert_file = '/home/super/.exec-server-cert.pem'
-    key_file = '/home/super/.exec-server-key.pem'
-    
-    if not os.path.exists(cert_file):
-        print('Generating self-signed cert...', file=sys.stderr)
-        subprocess.run([
-            'openssl', 'req', '-x509', '-newkey', 'rsa:2048',
-            '-keyout', key_file, '-out', cert_file,
-            '-days', '3650', '-nodes',
-            '-subj', '/CN=bl-exec-server'
-        ], check=True, capture_output=True)
-        os.chmod(key_file, 0o600)
-    
-    context = ssl.SSLContext(ssl.PROTOCOL_TLS_SERVER)
-    context.load_cert_chain(cert_file, key_file)
-    server.socket = context.wrap_socket(server.socket, server_side=True)
-    
-    print(f'Exec server listening on https://{args.host}:{args.port}/exec', file=sys.stderr)
-    print('Health check: https://<host>:<port>/health', file=sys.stderr)
-    
-    try:
-        server.serve_forever()
-    except KeyboardInterrupt:
-        print('\nShutting down', file=sys.stderr)
-
-if __name__ == '__main__':
-    main()

bin/exec-sign.sh

@@ -1,55 +0,0 @@
-#!/bin/bash
-# exec-sign.sh — call the bl exec-constrained server with SSH-signature auth.
-# No bearer token, no secret crosses the wire: you sign the request envelope
-# with your registered fleet key and the server verifies it against the
-# signers file. A signature is not a secret, so this never trips
-# secret-handling guardrails.
-#
-# Server: exec-constrained.py — named ops ONLY, no arbitrary shell.
-# Signature namespace: exec-constrained
-# Envelope: {"op","args","ts","nonce"} — op must be in the server allowlist:
-#   dm.send, dm.thread, dm.read, job.run, chat.messages, chat.send,
-#   health.check, exec.ping
-# Nonce: 16+ hex chars, replay-protected server-side. ts: unix epoch, ±300s skew.
-#
-# Usage: exec-sign.sh <op> '<args-json>' [identity] [keyfile] [url]
-#   op         a named op, e.g. exec.ping
-#   args-json  JSON object of the op's arguments, e.g. '{}'
-#   identity   defaults to operator-646 (must be a principal in the
-#              server's signers file)
-#   keyfile    defaults to ~/.ssh/id_frontdoor
-#   url        defaults to https://exec.muse-dev.online/exec (cloudflared).
-#              NOTE: the old VM Caddy /exec route to bl:8443 died with
-#              exec-server.py — do not point this at the sslip.io URL.
-set -euo pipefail
-OP="${1:?usage: exec-sign.sh <op> '<args-json>' [identity] [keyfile] [url]}"
-# NOTE: do NOT write this as ${2:-{}} — bash matches the first } as the
-# expansion's close brace and appends a literal } when $2 is set.
-ARGS_JSON="${2-}"
-if [ -z "$ARGS_JSON" ]; then ARGS_JSON='{}'; fi
-IDENTITY="${3:-operator-646}"
-KEY="${4:-$HOME/.ssh/id_frontdoor}"
-URL="${5:-https://exec.muse-dev.online/exec}"
-
-TS=$(date +%s)
-NONCE=$(python3 -c "import secrets; print(secrets.token_hex(16))")
-PAYLOAD=$(python3 -c "
-import json, sys
-op, args_json, ts, nonce = sys.argv[1:5]
-args = json.loads(args_json)
-if not isinstance(args, dict):
-    raise SystemExit('args-json must be a JSON object')
-print(json.dumps({'op': op, 'args': args, 'ts': int(ts), 'nonce': nonce}))
-" "$OP" "$ARGS_JSON" "$TS" "$NONCE")
-SIG=$(printf '%s' "$PAYLOAD" | ssh-keygen -Y sign -f "$KEY" -n exec-constrained)
-BODY=$(python3 -c "
-import json, sys
-ident, payload, sig = sys.argv[1:4]
-print(json.dumps({'identity': ident, 'payload': payload, 'signature': sig}))
-" "$IDENTITY" "$PAYLOAD" "$SIG")
-# Cloudflare Bot Fight Mode blocks python-urllib POSTs (error 1010):
-# use curl with a browser User-Agent instead.
-curl -sS -X POST "$URL" \
-  -H 'Content-Type: application/json' \
-  -H 'User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36' \
-  --data "$BODY"

bin/exec-watch.sh

@@ -1,59 +0,0 @@
-#!/bin/bash
-# exec-watch.sh — 15-min watchdog for the bl exec server's signature-auth path.
-# Signs a canary op as the exec-canary identity and POSTs it to the
-# local exec-constrained server. Records result in /home/super/.exec-watch-status.json
-# and appends to logs/exec-watch.log. Failures are pull-based (status file +
-# log); wire push alerting here if the fleet wants paging.
-#
-# Runs via systemd user timer exec-watch.timer (OnUnitActiveSec=15min).
-# Server: exec-constrained.py — named ops ONLY, no arbitrary shell;
-# signature namespace: exec-constrained, with {op,args,ts,nonce} envelope.
-# NOTE: signers file (/home/super/.exec-signers) is synced MANUALLY from the
-# VM's /srv/board/allowed_signers on identity renames/adds — bl cannot ssh
-# back to the VM, so there is no pull sync. Operator step, documented in
-# docs/TOKEN_POLICY.md.
-set -uo pipefail
-KEY=/home/super/.exec-canary
-URL=https://100.123.153.75:8444/exec
-STATUS=/home/super/.exec-watch-status.json
-LOG=/home/super/Projects/NetVM/logs/exec-watch.log
-
-ts=$(date +%s)
-nonce=$(python3 -c "import secrets; print(secrets.token_hex(16))")
-payload=$(python3 -c "import json,sys; print(json.dumps({'op':'exec.ping','args':{},'ts':int(sys.argv[1]),'nonce':sys.argv[2]}))" "$ts" "$nonce")
-sig=$(printf '%s' "$payload" | ssh-keygen -Y sign -f "$KEY" -n exec-constrained 2>/dev/null)
-if [ -z "${sig:-}" ]; then
-    result="sign-failed"
-else
-    body=$(python3 -c "import json,sys; print(json.dumps({'identity': 'exec-canary', 'payload': sys.argv[1], 'signature': sys.argv[2]}))" "$payload" "$sig")
-    out=$(curl -sk -m 25 -X POST "$URL" -H 'Content-Type: application/json' -d "$body" 2>/dev/null)
-    if echo "$out" | grep -q '"rc": 0'; then
-        result="ok"
-    else
-        result="bad-response"
-    fi
-fi
-
-now_iso=$(date -u +%FT%TZ)
-{
-python3 - "$STATUS" "$now_iso" "$result" <<'PYEOF'
-import json, sys
-status_path, now_iso, result = sys.argv[1], sys.argv[2], sys.argv[3]
-try:
-    st = json.load(open(status_path))
-except Exception:
-    st = {}
-if result == "ok":
-    st.update({"last_ok": now_iso, "last_fail": None,
-               "consecutive_failures": 0, "result": "ok"})
-else:
-    st.update({"last_ok": st.get("last_ok"),
-               "last_fail": now_iso,
-               "consecutive_failures": st.get("consecutive_failures", 0) + 1,
-               "result": result})
-json.dump(st, open(status_path, "w"))
-print(f"[{now_iso}] exec-watch: {result} "
-      f"(consecutive_failures={st['consecutive_failures']})")
-PYEOF
-} >> "$LOG" 2>&1
-[ "$result" = "ok" ]

bin/flap-check.sh

@@ -1,8 +0,0 @@
-#!/bin/bash
-# flap-check.sh - count browser relaunches per profile in the last hour
-# Called by the browser-flap-detector cron. Avoids nested SSH quoting hell.
-cutoff=$(date -u -d "1 hour ago" +%Y-%m-%dT%H:%M:%S)
-for prof in muse pip 646 opm; do
-  n=$(grep "\[$prof\]" /home/super/Projects/NetVM/chromebox-watchdog.log 2>/dev/null | grep "relaunch OK" | awk -v d="$cutoff" '{ts=substr($1,2,19); if (ts > d) c++} END {print c+0}')
-  echo "$prof:$n"
-done

bin/fleet-alert-check.sh

@@ -1,448 +0,0 @@
-#!/bin/bash
-# fleet-alert-check.sh — bl-side critical-condition detector for the fleet alerting pipeline.
-#
-# Closes the watchdog gap: agent-health.sh logs CRITICAL and restarts browsers,
-# but NOTHING pages anyone. This script detects critical conditions, counts
-# CONSECUTIVE failures, and emits alert records to an outbox that the
-# container-side fleet-alert-relay hook picks up and pages to #lobby.
-#
-# Checks (bl-side only; VM-side checks live in the container relay):
-#   cdp:<node>   headless Chromium CDP port not listening in the node's netns
-#                (catches zombie browsers: process alive, CDP not bound)
-#
-# Paging policy (env-overridable defaults — adjustable, not gates):
-#   FLEET_ALERT_THRESHOLD=2      consecutive failures before first page (~10 min at 5-min cadence)
-#   FLEET_ALERT_REALERT_MIN=30   re-page while still critical, at most every 30 min
-#   FLEET_ALERT_QUIET_HOURS=""   e.g. "23:00-07:00" (bl local time); empty = page 24/7.
-#                                First alert for a NEW incident always pages;
-#                                quiet hours only suppress re-pages.
-#   FLEET_ALERT_DRY_RUN=1        evaluate + print, write no state/outbox, no notify
-#   FLEET_ALERT_INJECT_FAIL=     test hook: comma-separated condition ids to force-fail
-#                                (e.g. FLEET_ALERT_INJECT_FAIL=cdp:pip)
-#   FLEET_BL_RELAY=1             re-enable the bl-side #lobby relay (default 0/off:
-#                                the container-side hook is the live pager; running
-#                                both double-posts every alert — 2026-10-06)
-#
-# State:  ~/.local/share/fleet-alert/state.json   (per-condition consecutive counters)
-# Outbox: ~/.local/share/fleet-alert/outbox.jsonl (ALERT/RECOVERY records for the relay)
-# Log:    /tmp/fleet-alert-check.log
-#
-# Alert delivery legs:
-#   1. outbox record  -> container relay -> signed #lobby post (primary page)
-#   2. best-effort `box-ctl.py notify` to currently-healthy agents (DM path needs a
-#      working browser; failures are logged, never fatal)
-#
-# Installed as user timer fleet-alert-check.timer (every 5 min), mirroring agent-health.timer.
-set -uo pipefail
-
-THRESHOLD="${FLEET_ALERT_THRESHOLD:-2}"
-REALERT_MIN="${FLEET_ALERT_REALERT_MIN:-30}"
-# Approval/input-wait TTLs (seconds): conditions failing longer than this are
-# auto-expired (input waits dismissed, key requests denied) instead of paging
-# forever. Overridable per environment.
-INPUT_WAIT_TTL="${FLEET_ALERT_INPUT_WAIT_TTL:-1800}"
-BROWSER_APPROVAL_TTL="${FLEET_ALERT_BROWSER_APPROVAL_TTL:-1800}"
-QUIET_HOURS="${FLEET_ALERT_QUIET_HOURS:-}"
-DRY_RUN="${FLEET_ALERT_DRY_RUN:-0}"
-INJECT_FAIL="${FLEET_ALERT_INJECT_FAIL:-}"
-STATE_DIR="${FLEET_ALERT_STATE_DIR:-$HOME/.local/share/fleet-alert}"
-BIN="$(cd "$(dirname "$0")" && pwd)"
-STATE="$STATE_DIR/state.json"
-OUTBOX="$STATE_DIR/outbox.jsonl"
-LOG="/tmp/fleet-alert-check.log"
-
-mkdir -p "$STATE_DIR"
-[ -f "$STATE" ] || echo '{}' > "$STATE"
-touch "$OUTBOX"
-
-NOW=$(date +%s)
-log() { echo "$(date -Iseconds) $*" >> "$LOG"; }
-
-# --- shared consecutive-failure state machine (also used by the container relay) ---
-# usage: state_machine <cond> <failing 0|1> [ttl_seconds] -> prints "<ACTION> <fails>"
-# When ttl_seconds > 0 and the condition has failed longer than the TTL,
-# prints "EXPIRED <fails>" so the caller can auto-resolve (dismiss/deny).
-# State entries track first_fail_ts (epoch of first consecutive failure).
-# ACTION: ALERT_FIRST | ALERT_REALERT | RECOVERY | SUPPRESSED | EXPIRED | NONE
-state_machine() {
-    local cond="$1" failing="$2" ttl="${3:-0}"
-    THRESHOLD="$THRESHOLD" REALERT_MIN="$REALERT_MIN" QUIET_HOURS="$QUIET_HOURS" \
-    FLEET_ALERT_DRY_RUN="$DRY_RUN" python3 - "$STATE" "$cond" "$failing" "$ttl" <<'PYEOF'
-import json, os, sys, time
-state_path, cond, failing_s = sys.argv[1], sys.argv[2], sys.argv[3]
-ttl_seconds = int(sys.argv[4]) if len(sys.argv) > 4 else 0
-failing = failing_s == "1"
-threshold = int(os.environ.get("THRESHOLD", "2"))
-realert_min = int(os.environ.get("REALERT_MIN", "30"))
-qh = os.environ.get("QUIET_HOURS", "")
-dry = os.environ.get("FLEET_ALERT_DRY_RUN") == "1"
-now = int(time.time())
-def in_quiet(spec):
-    if not spec:
-        return False
-    try:
-        a, b = spec.split("-")
-        def m(s):
-            h, mi = s.split(":")
-            return int(h) * 60 + int(mi)
-        cur = time.localtime().tm_hour * 60 + time.localtime().tm_min
-        s, e = m(a), m(b)
-        return (s <= cur < e) if s <= e else (cur >= s or cur < e)
-    except Exception:
-        return False
-try:
-    st = json.load(open(state_path))
-except Exception:
-    st = {}
-e = st.get(cond) or {"fails": 0, "alerted": False, "last_alert_ts": 0}
-action = "NONE"
-if failing:
-    if int(e.get("fails", 0)) == 0:
-        e["first_fail_ts"] = now
-    e["fails"] = int(e.get("fails", 0)) + 1
-    # TTL expiry: failing longer than ttl_seconds -> EXPIRED (caller auto-resolves)
-    if ttl_seconds > 0 and now - int(e.get("first_fail_ts", now)) >= ttl_seconds:
-        action = "EXPIRED"
-        # Reset so a fresh incident starts clean after the caller resolves it
-        e["fails"] = 0
-        e["alerted"] = False
-        e.pop("first_fail_ts", None)
-    else:
-        due = e["fails"] >= threshold and (
-            not e.get("alerted") or now - int(e.get("last_alert_ts", 0)) >= realert_min * 60
-        )
-        if due:
-            first = not e.get("alerted")
-            if not first and in_quiet(qh):
-                action = "SUPPRESSED"
-            else:
-                action = "ALERT_FIRST" if first else "ALERT_REALERT"
-                e["alerted"] = True
-                e["last_alert_ts"] = now
-else:
-    if e.get("alerted"):
-        action = "RECOVERY"
-    e["fails"] = 0
-    e["alerted"] = False
-    e.pop("first_fail_ts", None)
-st[cond] = e
-if not dry:
-    json.dump(st, open(state_path, "w"))
-print(action, e["fails"])
-PYEOF
-}
-
-emit_record() { # kind cond detail consecutive
-    local kind="$1" cond="$2" detail="$3" consec="$4"
-    local id
-    id=$(tr -d '-' < /proc/sys/kernel/random/uuid | cut -c1-12)
-    local rec
-    rec=$(python3 -c 'import json,sys; print(json.dumps({"id":sys.argv[1],"ts":int(sys.argv[2]),"kind":sys.argv[3],"source":"bl","condition":sys.argv[4],"detail":sys.argv[5],"consecutive":int(sys.argv[6])}))' \
-        "$id" "$NOW" "$kind" "$cond" "$detail" "$consec")
-    if [ "$DRY_RUN" = "1" ]; then
-        log "DRY-RUN would emit: $rec"
-    else
-        echo "$rec" >> "$OUTBOX"
-        log "emitted $kind $cond x$consec"
-    fi
-}
-
-box_notify() {
-    # Best-effort DM to healthy agents via box-ctl. Never fatal.
-    # Fan-out runs in parallel with a per-notify timeout so one hung DM
-    # path can't stall the 5-minute check loop.
-    local cond="$1"
-    local detail="$2"
-    local msg="[fleet-alert] CRITICAL ${cond}: ${detail}"
-    msg="${msg:0:240}"
-    local agent pids=""
-    for agent in $HEALTHY_AGENTS; do
-        if [ "$DRY_RUN" = "1" ]; then
-            log "DRY-RUN would notify $agent"
-            continue
-        fi
-        (
-            if timeout 60 python3 "$BIN/box-ctl.py" notify "$agent" "$msg" >/dev/null 2>&1; then
-                log "notified $agent re $cond"
-            else
-                log "notify $agent failed re $cond (best-effort)"
-            fi
-        ) &
-        pids="$pids $!"
-    done
-    local p
-    for p in $pids; do wait "$p" 2>/dev/null; done
-}
-
-notify_input_wait() {
-    # Targeted DM for input waits (2026-10-05): DM ONLY the specific agent
-    # whose session is waiting for human input -- not a broadcast to all
-    # healthy agents. The #lobby post still fires via the relay leg for
-    # human visibility; this DM ensures the responsible operator sees it
-    # in their sidechat without digging through lobby noise.
-    # Best-effort: never fatal to the 5-minute check loop.
-    local node="$1"
-    local detail="$2"
-    local msg="[fleet-alert] INPUT WAIT: ${detail} -- reply: box approval reply ${node} \"<msg>\" or box notify ${node} \"<msg>\""
-    msg="${msg:0:900}"
-    if [ "$DRY_RUN" = "1" ]; then
-        log "DRY-RUN would DM $node re input_wait"
-        return 0
-    fi
-    if timeout 60 python3 "$BIN/box-ctl.py" notify "$node" "$msg" >/dev/null 2>&1; then
-        log "input_wait targeted DM sent to $node"
-    else
-        log "input_wait DM to $node failed (best-effort, non-fatal)"
-    fi
-}
-
-injected() { # cond -> 0 if injected-fail
-    case ",$INJECT_FAIL," in *,"$1,"*) return 0;; *) return 1;; esac
-}
-
-HEALTHY_AGENTS=""
-
-"$BIN/netvm-registry.py" 2>/dev/null | while IFS=: read -r node port; do
-    [ -n "$node" ] && [ -n "$port" ] || continue
-    cond="cdp:$node"
-    procs=$(pgrep -f "chromium.*profiles/$node" 2>/dev/null | wc -l)
-    if sudo -n ip netns exec "warp-$node" ss -tln 2>/dev/null | grep -q ":$port "; then
-        failing=0
-        HEALTHY_AGENTS="$HEALTHY_AGENTS $node"
-    else
-        failing=1
-    fi
-    injected "$cond" && failing=1
-    detail="CDP $port not listening in netns warp-$node (chromium procs=$procs)"
-    # NOTE: HEALTHY_AGENTS set inside the pipeline subshell is lost; recompute below.
-    read -r action fails < <(state_machine "$cond" "$failing")
-    case "$action" in
-        ALERT_FIRST|ALERT_REALERT)
-            emit_record "ALERT" "$cond" "$detail" "$fails"
-            echo "$cond" >> "$STATE_DIR/.alerts.tmp"
-            ;;
-        RECOVERY)
-            emit_record "RECOVERY" "$cond" "CDP $port listening again in netns warp-$node" "$fails"
-            ;;
-        SUPPRESSED)
-            log "$cond still critical x$fails — re-page suppressed by quiet hours ($QUIET_HOURS)"
-            ;;
-    esac
-done
-
-# --- Agent approval blockage & input wait detection ---
-"$BIN/netvm-registry.py" 2>/dev/null | while IFS=: read -r node port; do
-    [ -n "$node" ] || continue
-    node_data=$(python3 -c "
-import sys, json
-sys.path.insert(0, '$BIN')
-import approvals
-info = approvals.inspect_node_approvals('$node')
-out = {
-    'has_pending': info.get('has_pending', False),
-    'target': info.get('target') or info.get('ip') or 'unknown',
-    'title': info.get('title') or '',
-    'waits': info.get('input_waits') or []
-}
-print(json.dumps(out))
-" 2>/dev/null || echo '{"has_pending":false,"target":"unknown","title":"","waits":[]}')
-
-    # 1. Egress permission dialog
-    cond="approval:$node"
-    has_pending=$(python3 -c "import json,sys; print(1 if json.loads(sys.argv[1]).get('has_pending') else 0)" "$node_data" 2>/dev/null || echo 0)
-    if [ "$has_pending" = "1" ]; then
-        failing=1
-        target=$(python3 -c "import json,sys; print(json.loads(sys.argv[1]).get('target','unknown'))" "$node_data" 2>/dev/null || echo unknown)
-        detail="Agent $node held up on browser approval for $target"
-    else
-        failing=0
-        detail="Agent $node approvals clear"
-    fi
-    injected "$cond" && failing=1
-    read -r action fails < <(state_machine "$cond" "$failing" "$BROWSER_APPROVAL_TTL")
-    case "$action" in
-        ALERT_FIRST|ALERT_REALERT)
-            emit_record "ALERT" "$cond" "$detail" "$fails"
-            echo "$cond" >> "$STATE_DIR/.alerts.tmp"
-            ;;
-        RECOVERY)
-            emit_record "RECOVERY" "$cond" "$detail" "$fails"
-            ;;
-        SUPPRESSED)
-            log "$cond still critical x$fails — re-page suppressed"
-            ;;
-        EXPIRED)
-            # Browser approval dialog exceeded BROWSER_APPROVAL_TTL without a
-            # human decision: fail closed by denying it.
-            log "$cond EXPIRED after ${BROWSER_APPROVAL_TTL}s without human decision — auto-denying (fail closed)"
-            python3 - "$node" <<'PYEOF3'
-import sys, json
-sys.path.insert(0, "/home/super/Projects/NetVM/bin")
-import approvals
-node = sys.argv[1]
-print(json.dumps(approvals.deny_node_approval(node, caller="approval-ttl-expire")))
-approvals.log_box_ctl("approval-expired", name=node, caller="approval-ttl-expire",
-                      extra={"note": "browser approval TTL elapsed; auto-denied (fail closed)"})
-PYEOF3
-            emit_record "RECOVERY" "$cond" "Agent $node browser approval expired after ${BROWSER_APPROVAL_TTL}s; auto-denied" "$fails"
-            ;;
-    esac
-
-    # 2. Sidebar task waiting on human input
-    cond_in="input_wait:$node"
-    wait_summary=$(python3 -c "
-import json,sys
-w = json.loads(sys.argv[1]).get('waits', [])
-if w:
-    print('; '.join(f\"{item.get('task')}: {item.get('status')}\" for item in w)[:120])
-" "$node_data" 2>/dev/null || true)
-
-    if [ -n "$wait_summary" ]; then
-        failing_in=1
-        detail_in="Agent $node task waiting for human input: $wait_summary"
-    else
-        failing_in=0
-        detail_in="Agent $node tasks running"
-    fi
-    injected "$cond_in" && failing_in=1
-    read -r action_in fails_in < <(state_machine "$cond_in" "$failing_in" "$INPUT_WAIT_TTL")
-    case "$action_in" in
-        ALERT_FIRST|ALERT_REALERT)
-            emit_record "ALERT" "$cond_in" "$detail_in" "$fails_in"
-            echo "$cond_in|$detail_in" >> "$STATE_DIR/.alerts.tmp"
-            ;;
-        RECOVERY)
-            emit_record "RECOVERY" "$cond_in" "$detail_in" "$fails_in"
-            ;;
-        SUPPRESSED)
-            log "$cond_in still critical x$fails_in — re-page suppressed"
-            ;;
-        EXPIRED)
-            # Input wait exceeded INPUT_WAIT_TTL without human response:
-            # auto-dismiss so the agent unblocks. Log the expiry and emit a
-            # RECOVERY record (the wait is gone, not merely un-paged).
-            log "$cond_in EXPIRED after ${INPUT_WAIT_TTL}s without human input — auto-dismissing"
-            python3 - "$node" <<'PYEOF2'
-import sys
-sys.path.insert(0, "/home/super/Projects/NetVM/bin")
-import approvals, json
-node = sys.argv[1]
-info = approvals.inspect_node_approvals(node)
-for w in info.get("input_waits", []) or []:
-    t = w.get("task")
-    if t:
-        approvals.mark_wait_responded(node, t, caller="approval-ttl-expire")
-approvals.log_box_ctl("approval-wait-expired", name=node, caller="approval-ttl-expire",
-                      extra={"note": "input wait TTL elapsed; auto-dismissed"})
-print(json.dumps(approvals.dismiss_node_task(node, caller="approval-ttl-expire")))
-PYEOF2
-            emit_record "RECOVERY" "$cond_in" "Agent $node input wait expired after ${INPUT_WAIT_TTL}s; auto-dismissed" "$fails_in"
-            ;;
-    esac
-done
-
-# --- Warp partition detection (2026-10-04) ---
-# A partitioned node has a live browser + CDP but no internet egress: the
-# chromebox watchdog sees a healthy browser while all automation fails.
-# Condition id: partition:<node>. The detail names the node, the WireGuard
-# handshake age, the egress probe result, and the timestamp, and says
-# PARTITION explicitly so #lobby readers can tell a network partition from
-# a browser crash at a glance. Anti-spam comes from the shared consecutive-
-# failure state machine (2 consecutive failures before first page, re-page
-# at most every 30 min).
-WARP_PROBE_URL="${WARP_PROBE_URL:-https://1.1.1.1/cdn-cgi/trace}"
-warp_partition_probe() { # <node> -> prints "<handshake_age_s|unknown> <ok|FAIL>"
-    local node="$1" iface epoch now age_s
-    iface=$(sudo -n ip netns exec "warp-$node" sh -c 'wg show interfaces 2>/dev/null | head -1')
-    now=$(date +%s)
-    epoch=$(sudo -n ip netns exec "warp-$node" wg show "$iface" latest-handshakes 2>/dev/null | awk '{print $2}')
-    case "$epoch" in ''|*[!0-9]*) age_s="unknown" ;; *) age_s=$(( now - epoch )) ;; esac
-    if sudo -n ip netns exec "warp-$node" curl -s -m 8 -o /dev/null "$WARP_PROBE_URL" 2>/dev/null; then
-        echo "$age_s ok"
-    else
-        echo "$age_s FAIL"
-    fi
-}
-
-"$BIN/netvm-registry.py" 2>/dev/null | while IFS=: read -r node port; do
-    [ -n "$node" ] && [ -n "$port" ] || continue
-    cond="partition:$node"
-    read -r hs_age probe_res < <(warp_partition_probe "$node")
-    ts=$(date -u +%FT%TZ)
-    if [ "$probe_res" = "ok" ]; then
-        failing=0
-        detail="warp egress restored for $node at $ts (probe $WARP_PROBE_URL ok)"
-    else
-        failing=1
-        detail="PARTITION $node: warp egress down at $ts (handshake ${hs_age}s ago, probe $WARP_PROBE_URL FAILED)"
-    fi
-    if injected "$cond"; then
-        failing=1
-        detail="PARTITION $node: warp egress down at $ts (handshake ${hs_age}s ago, probe $WARP_PROBE_URL FAILED) [INJECTED]"
-    fi
-    read -r action fails < <(state_machine "$cond" "$failing")
-    case "$action" in
-        ALERT_FIRST|ALERT_REALERT)
-            emit_record "ALERT" "$cond" "$detail" "$fails"
-            echo "$cond" >> "$STATE_DIR/.alerts.tmp"
-            ;;
-        RECOVERY)
-            emit_record "RECOVERY" "$cond" "$detail" "$fails"
-            ;;
-        SUPPRESSED)
-            log "$cond still critical x$fails — re-page suppressed by quiet hours ($QUIET_HOURS)"
-            ;;
-    esac
-done
-
-# Recompute healthy agents in the main shell (pipeline subshell above can't export).
-HEALTHY_AGENTS=""
-"$BIN/netvm-registry.py" 2>/dev/null | while IFS=: read -r node port; do
-    [ -n "$node" ] && [ -n "$port" ] || continue
-    if sudo -n ip netns exec "warp-$node" ss -tln 2>/dev/null | grep -q ":$port "; then
-        echo "$node"
-    fi
-done > "$STATE_DIR/.healthy.tmp"
-HEALTHY_AGENTS=$(tr '\n' ' ' < "$STATE_DIR/.healthy.tmp")
-rm -f "$STATE_DIR/.healthy.tmp"
-
-# Notify for this run's alerts (best effort). Skip entirely when nothing is healthy
-# (notify needs a working browser via dm.py) or in dry-run.
-if [ -n "$HEALTHY_AGENTS" ] && [ -f "$STATE_DIR/.alerts.tmp" ]; then
-    while IFS= read -r line; do
-        # alerts.tmp format: "cond" or "cond|detail" (input_wait carries detail)
-        cond="${line%%|*}"
-        detail="${line#*|}"
-        [ "$detail" = "$line" ] && detail=""
-        [ -n "$cond" ] || continue
-        case "$cond" in
-            input_wait:*)
-                # Targeted: DM only the waiting agent, not a broadcast.
-                node="${cond#input_wait:}"
-                if [ -n "$detail" ]; then
-                    notify_input_wait "$node" "$detail"
-                else
-                    box_notify "$cond" "see #lobby for detail"
-                fi
-                ;;
-            *)
-                box_notify "$cond" "see #lobby for detail"
-                ;;
-        esac
-    done < "$STATE_DIR/.alerts.tmp"
-elif [ -f "$STATE_DIR/.alerts.tmp" ]; then
-    log "no healthy agents — box notify skipped (DM path needs a working browser)"
-fi
-rm -f "$STATE_DIR/.alerts.tmp"
-
-tail -500 "$LOG" > "$LOG.tmp" 2>/dev/null && mv "$LOG.tmp" "$LOG"
-log "check complete"
-
-# Bl-side #lobby relay: DISABLED by default (FLEET_BL_RELAY=1 to re-enable).
-# The container-side hook is the live pager; the bl relay never successfully
-# posted (missing CHAT_KEYFILE) and enabling it now would double-post every
-# alert in a second format. Re-enable only alongside retiring the container
-# hook (and per the relay header, with opm sign-off).
-if [ "${FLEET_BL_RELAY:-0}" = "1" ] && [ "$DRY_RUN" -eq 0 ] && [ -x "$BIN/fleet-alert-relay.sh" ]; then
-    "$BIN/fleet-alert-relay.sh" >> "$LOG" 2>&1 || true
-fi

bin/fleet-alert-relay.sh

@@ -1,336 +0,0 @@
-#!/bin/bash
-# bin/fleet-alert-relay.sh — idempotent relay: fleet-alert outbox.jsonl -> #lobby.
-#
-# PROPOSAL ONLY (board ticket ae28ac8b735f). No bl infra is touched by this
-# branch; deployment needs opm review + sign-off. See
-# docs/FLEET-ALERT-DUP-POST-GATE.md.
-#
-# NOTE (2026-10-06): auto-invoke from fleet-alert-check.sh is disabled by
-# default (FLEET_BL_RELAY=1 re-enables). The container-side hook pages
-# #lobby today; do not re-enable without retiring it first.
-#
-# The 2026-10-05 11:28Z incident: one RECOVERY record in the outbox became two
-# identical verified #lobby posts (seq 642/643, 3.35s apart) because the relay
-# leg had no idempotency: append-only outbox, no consume tracking, no content
-# dedup, and the chat POST API has no idempotency key.
-#
-# Gates implemented here:
-#   1. posted-watermark — each posted outbox record id is appended to
-#      posted.log; records already posted are skipped.
-#   2. content-hash + 10-min TTL — sha256 of the formatted alert text in a
-#      file-backed seen-set (seen-hashes.log); identical text re-posts inside
-#      the window are dropped. Catches same-record re-posts AND identical
-#      text from different records.
-#   3. retry discipline — never blind-retry a chat POST after a timeout /
-#      phantom-000. Read the #lobby tail first; re-post only if absent.
-#
-# Usage:
-#   fleet-alert-relay.sh [--dry-run] [--self-test]
-#
-# Env: FLEET_ALERT_DIR (default ~/.local/share/fleet-alert),
-#      CHAT_BASE (default https://chat.muse-dev.online),
-#      CHAT_IDENTITY (default operator-646), CHAT_KEYFILE (default ~/.ssh/id_frontdoor)
-set -euo pipefail
-
-ALERT_DIR="${FLEET_ALERT_DIR:-$HOME/.local/share/fleet-alert}"
-OUTBOX="$ALERT_DIR/outbox.jsonl"
-POSTED="$ALERT_DIR/posted.log"
-SEEN="$ALERT_DIR/seen-hashes.log"
-LOCKF="$ALERT_DIR/relay.lock"
-
-CHAT_BASE="${CHAT_BASE:-https://chat.muse-dev.online}"
-API="$CHAT_BASE/api/chat"
-IDENTITY="${CHAT_IDENTITY:-operator-646}"
-KEYFILE="${CHAT_KEYFILE:-$HOME/.ssh/id_frontdoor}"
-CHANNEL="#lobby"
-TTL_SECS=600
-UA='Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0 Safari/537.36'
-
-DRY_RUN=0
-log() { echo "[fleet-alert-relay] $*" >&2; }
-
-# ---------------------------------------------------------------- transport
-# Factored so --self-test can stub them. Return contracts:
-#   transport_tail -> prints "<ts> <identity> <message>" lines for recent #lobby (one per line, tab-safe: message may contain spaces, fields split on first two spaces)
-#   transport_post "<text>" -> prints "OK <msg-id>" or "UNKNOWN <reason>"; exit 0 always (callers decide)
-
-transport_tail() {
-  curl -s -m 15 -A "$UA" "$API/history?channel=%23lobby&limit=50" \
-    | python3 -c '
-import json,sys
-try:
-    d=json.load(sys.stdin)
-except Exception:
-    sys.exit(0)
-for m in d.get("messages",[]):
-    msg=(m.get("message") or "").replace(chr(10)," ")
-    print(m.get("ts"), m.get("identity"), msg)
-'
-}
-
-transport_post() { # $1 = text
-  local text="$1" ts sig payload resp http
-  ts="$(date +%s)"
-  if ! sig="$(sign_payload "$(printf '%s\n%s\n%s' "$ts" "$CHANNEL" "$text")")"; then
-    echo "UNKNOWN sign-failed($KEYFILE)"; return 0
-  fi
-  payload="$(MSG="$text" TS="$ts" SIG="$sig" python3 -c '
-import json,os
-print(json.dumps({"channel":os.environ["CHANNEL"],"identity":os.environ["IDENTITY"],
-  "message":os.environ["MSG"],"ts":int(os.environ["TS"]),"signature":os.environ["SIG"]}))')"
-  resp="$(curl -s -m 20 -A "$UA" -X POST -H 'Content-Type: application/json' \
-    -d "$payload" -w '\n%{http_code}' "$API/post" 2>/dev/null || echo -e '\n000')"
-  http="$(printf '%s' "$resp" | tail -n1)"
-  resp="$(printf '%s' "$resp" | sed '$d')"
-  case "$http" in
-    2*)
-      local mid
-      mid="$(printf '%s' "$resp" | python3 -c '
-import json,sys
-try:
-    d=json.load(sys.stdin); print(d.get("id","") if d.get("ok") else "")
-except Exception:
-    print("")')"
-      if [ -n "$mid" ]; then echo "OK $mid"; else echo "UNKNOWN bad-body-$http"; fi
-      ;;
-    *) echo "UNKNOWN http-$http" ;;
-  esac
-  return 0
-}
-
-# file-based signing (never pipe): stdin-piped signatures intermittently
-# fail verification (2026-10-02 incident).
-sign_payload() { # $1 = payload -> armored signature
-  local tmpd pf
-  tmpd="$(mktemp -d)" || return 1
-  pf="$tmpd/payload"
-  printf '%s' "$1" > "$pf"
-  rm -f "$pf.sig"
-  ssh-keygen -Y sign -f "$KEYFILE" -n chat "$pf" </dev/null >/dev/null 2>&1 \
-    || { rm -rf "$tmpd"; return 1; }
-  cat "$pf.sig"
-  rm -rf "$tmpd"
-}
-
-# ----------------------------------------------------------------- formatting
-# Canonical text for one outbox record (JSON on stdin):
-#   ALERT    -> [fleet-alert] CRITICAL: <node> warp PARTITION x<n>
-#   RECOVERY -> [fleet-alert] RECOVERED: <node> warp PARTITION
-# condition is "<type>:<node>" (partition|cdp).
-format_text() {
-  python3 -c '
-import json,sys
-r=json.load(sys.stdin)
-cond=r.get("condition","")
-typ,_,node=cond.partition(":")
-label={"partition":"PARTITION","cdp":"CDP"}.get(typ,typ.upper())
-kind=r.get("kind","")
-if kind=="ALERT":
-    n=r.get("consecutive",r.get("count","?"))
-    print("[fleet-alert] CRITICAL: %s warp %s x%s" % (node,label,n))
-elif kind=="RECOVERY":
-    print("[fleet-alert] RECOVERED: %s warp %s" % (node,label))
-else:
-    print("[fleet-alert] %s: %s warp %s" % (kind,node,label))
-'
-}
-
-content_hash() { printf '%s' "$1" | sha256sum | cut -d' ' -f1; }
-
-# ------------------------------------------------------------- gate 1: watermark
-already_posted() { # $1 = record id
-  [ -f "$POSTED" ] && grep -qxF "$1" "$POSTED"
-}
-mark_posted() { # $1 = record id
-  printf '%s\n' "$1" >> "$POSTED"
-}
-
-# ------------------------------------------------------- gate 2: content-hash TTL
-seen_recently() { # $1 = hash -> 0 if seen within TTL
-  local h="$1" now ts
-  [ -f "$SEEN" ] || return 1
-  now="$(date +%s)"
-  while read -r h2 ts _rest; do
-    [ "$h2" = "$h" ] || continue
-    if [ "$(( now - ts ))" -lt "$TTL_SECS" ]; then return 0; fi
-  done < "$SEEN"
-  return 1
-}
-mark_seen() { # $1 = hash
-  printf '%s %s\n' "$1" "$(date +%s)" >> "$SEEN"
-}
-prune_seen() {
-  local now tmp
-  [ -f "$SEEN" ] || return 0
-  now="$(date +%s)"; tmp="$(mktemp)"
-  while read -r h ts _rest; do
-    [ -n "$h" ] && [ "$(( now - ts ))" -lt "$TTL_SECS" ] && printf '%s %s\n' "$h" "$ts" >> "$tmp"
-  done < "$SEEN"
-  mv "$tmp" "$SEEN"
-}
-
-# ------------------------------------------------- gate 3: check-before-(re)post
-# 0 if the exact text appears in the recent #lobby tail.
-lobby_has_text() { # $1 = text
-  local want="$1"
-  # transport_tail prints "<ts> <identity> <message>" per line; the message
-  # itself may contain spaces, so strip only the first two fields.
-  transport_tail | python3 -c '
-import sys
-want=sys.argv[1]
-for line in sys.stdin:
-    parts=line.rstrip("\n").split(" ",2)
-    if len(parts)==3 and parts[2]==want:
-        sys.exit(0)
-sys.exit(1)' "$want"
-}
-
-relay_record() { # $1 = record id, $2 = formatted text
-  local rid="$1" text="$2" h out
-  h="$(content_hash "$text")"
-
-  # Gate 1: posted watermark
-  if already_posted "$rid"; then
-    log "skip $rid: already in posted.log"
-    return 0
-  fi
-  # Gate 2: content hash within TTL
-  if seen_recently "$h"; then
-    log "skip $rid: identical text posted <10min ago (hash $h)"
-    mark_posted "$rid"   # consume it so it never retries later
-    return 0
-  fi
-  # Gate 3a: somebody else already posted it (covers agent-driven relay overlap)
-  if lobby_has_text "$text"; then
-    log "skip $rid: text already in #lobby tail"
-    mark_posted "$rid"; mark_seen "$h"
-    return 0
-  fi
-
-  if [ "$DRY_RUN" = 1 ]; then
-    log "dry-run: would post $rid: $text"
-    return 0
-  fi
-
-  # Attempt one POST. Any ambiguous outcome -> verify via tail, never blind-retry.
-  out="$(transport_post "$text")"
-  case "$out" in
-    OK*)
-      log "posted $rid -> ${out#OK }"
-      mark_posted "$rid"; mark_seen "$h"
-      return 0
-      ;;
-    UNKNOWN*)
-      log "post $rid outcome UNKNOWN (${out#UNKNOWN }); checking #lobby tail before any retry"
-      sleep 2
-      if lobby_has_text "$text"; then
-        log "post $rid landed despite ${out#UNKNOWN } — marking posted, no retry"
-        mark_posted "$rid"; mark_seen "$h"
-        return 0
-      fi
-      log "post $rid absent from tail — ONE retry only"
-      out="$(transport_post "$text")"
-      case "$out" in
-        OK*)
-          log "retry posted $rid -> ${out#OK }"
-          mark_posted "$rid"; mark_seen "$h"
-          return 0
-          ;;
-        *)
-          log "ERROR: post $rid still ${out%% *} after one retry; leaving unposted for next run (tail-check will catch it if it landed)"
-          return 1
-          ;;
-      esac
-      ;;
-  esac
-}
-
-main() {
-  local rid rec text fails=0
-  [ -f "$OUTBOX" ] || { log "no outbox ($OUTBOX); nothing to do"; return 0; }
-  mkdir -p "$ALERT_DIR"
-  touch "$POSTED" "$SEEN"
-  prune_seen
-  while IFS= read -r rec; do
-    [ -n "$rec" ] || continue
-    rid="$(printf '%s' "$rec" | python3 -c 'import json,sys; print(json.load(sys.stdin).get("id",""))')"
-    [ -n "$rid" ] || { log "skip record with no id"; continue; }
-    text="$(printf '%s' "$rec" | format_text)"
-    relay_record "$rid" "$text" || fails=$((fails+1))
-  done < "$OUTBOX"
-  return "$fails"
-}
-
-# ------------------------------------------------------------------ self-test
-# Acceptance: two identical alert submissions <10 min apart -> exactly one
-# #lobby post. Stubs the transport; exercises all three gates.
-# NOTE: transport_post is invoked via command substitution (subshell), so the
-# stub counts calls with a file, not a variable.
-self_test() {
-  local td calls lobby ok=1 n sk sig_out old_key
-  td="$(mktemp -d)"; export FLEET_ALERT_DIR="$td"
-  ALERT_DIR="$td"; OUTBOX="$td/outbox.jsonl"; POSTED="$td/posted.log"
-  SEEN="$td/seen-hashes.log"; LOCKF="$td/relay.lock"
-  calls="$td/calls.log"; lobby="$td/lobby.log"
-  touch "$calls" "$lobby"
-  # sign_payload must round-trip with a valid key and fail cleanly without
-  # one (2026-10-06: missing ~/.ssh/id_frontdoor broke every #lobby post
-  # with an undiagnosable bare "sign-failed").
-  old_key="$KEYFILE"
-  sk="$td/signkey"
-  ssh-keygen -t ed25519 -f "$sk" -N '' -q >/dev/null 2>&1 \
-    || { echo "FAIL: cannot generate ephemeral test key"; ok=0; }
-  if KEYFILE="$sk" sig_out="$(sign_payload "self-test")"; then
-    case "$sig_out" in
-      *"BEGIN SSH SIGNATURE"*) : ;;
-      *) echo "FAIL: sign_payload output not armored"; ok=0 ;;
-    esac
-  else
-    echo "FAIL: sign_payload failed with a valid key"; ok=0
-  fi
-  if KEYFILE="$td/no-such-key" sign_payload "self-test" >/dev/null 2>&1; then
-    echo "FAIL: sign_payload succeeded with a missing key"; ok=0
-  fi
-  KEYFILE="$old_key"
-  # Two identical submissions: same text, different record ids (the 11:28Z shape)
-  printf '%s\n' \
-    '{"id":"rec-A","ts":1791199616,"kind":"RECOVERY","condition":"partition:def"}' \
-    '{"id":"rec-B","ts":1791199617,"kind":"RECOVERY","condition":"partition:def"}' \
-    > "$OUTBOX"
-  transport_post() { # stub: 1st call "times out" but the server DID accept it
-    echo "call" >> "$calls"
-    printf '%s\n' "$1" >> "$lobby"
-    n=$(wc -l < "$calls")
-    if [ "$n" = 1 ]; then echo "UNKNOWN phantom-000"; else echo "OK stub-$n"; fi
-    return 0
-  }
-  transport_tail() { # stub: the lobby shows whatever "landed"
-    while IFS= read -r m; do printf '%s opm %s\n' "$(date +%s)" "$m"; done < "$lobby"
-  }
-  main
-  n=$(wc -l < "$calls")
-  [ "$n" -eq 1 ] || { echo "FAIL: expected exactly 1 post call, got $n"; ok=0; }
-  grep -qx "rec-A" "$POSTED" || { echo "FAIL: rec-A not watermarked"; ok=0; }
-  grep -qx "rec-B" "$POSTED" || { echo "FAIL: rec-B not watermarked (gate 2 should consume it)"; ok=0; }
-  # Identical re-submission (new record id, same text) must not post again
-  printf '%s\n' '{"id":"rec-C","ts":1791199700,"kind":"RECOVERY","condition":"partition:def"}' >> "$OUTBOX"
-  main
-  n=$(wc -l < "$calls")
-  [ "$n" -eq 1 ] || { echo "FAIL: identical re-submission posted again (calls=$n)"; ok=0; }
-  rm -rf "$td"
-  [ "$ok" = 1 ] && echo "SELF-TEST PASS: 2 identical submissions -> 1 post call; re-submission -> 0 new calls" && return 0
-  return 1
-}
-
-case "${1:-}" in
-  --dry-run) DRY_RUN=1; shift ;;
-  --self-test) self_test; exit $? ;;
-esac
-
-# Single-flight the relay (defense in depth; the detector's own flock is separate).
-exec 9>"$LOCKF"
-if ! flock -n 9; then
-  log "another relay run holds the lock; exiting"
-  exit 0
-fi
-main

bin/fleet-status.sh

@@ -1,12 +0,0 @@
-#!/bin/bash
-# fleet-status.sh - one-line health per profile: browser up/down + relay code
-declare -A ports=( [muse]=9410 [pip]=9420 [646]=9430 [opm]=9440 )
-declare -A veth=( [muse]=10.201.35.2 [pip]=10.201.87.2 [646]=10.201.202.2 [opm]=10.201.157.2 )
-for p in muse pip 646 opm; do
-  port=${ports[$p]}
-  if pgrep -f "remote-debugging-port=$port" >/dev/null 2>&1; then b=UP; else b=DOWN; fi
-  r=$(curl -s -m 4 -o /dev/null -w "%{http_code}" "http://${veth[$p]}:$port/json/version" 2>/dev/null || echo 000)
-  echo "$p: browser=$b relay=$r"
-done
-echo ===
-tail -20 /home/super/Projects/NetVM/chromebox-watchdog.log | grep -E "FAILED|unhealthy" | tail -4

bin/followup-sweeper.py

@@ -1,578 +0,0 @@
-#!/usr/bin/env python3
-"""
-followup-sweeper.py — Autonomous follow-up deadline tracking and nudge sweeper.
-
-Monitors pending follow-ups in followups.json, delivers progressive nudges to
-recipients when deadlines expire (in-thread first, Main Chat on final nudge),
-and executes terminal escalations to opm when all nudges are exhausted.
-
-Usage:
-    python3 followup-sweeper.py --once
-    python3 followup-sweeper.py --loop --interval 30
-"""
-
-import argparse
-import json
-import os
-import re
-import subprocess
-import sys
-import time
-from datetime import datetime, timezone, timedelta
-from pathlib import Path
-
-# Paths
-NETVM_ROOT = Path("/home/super/Projects/NetVM")
-BIN_DIR = NETVM_ROOT / "bin"
-JOBS_DIR = NETVM_ROOT / "jobs"
-FOLLOWUPS_FILE = NETVM_ROOT / "followups.json"
-JOB_LOG = NETVM_ROOT / "job-log.jsonl"
-DM_LOG = NETVM_ROOT / "dm-log.jsonl"
-DM_PY = BIN_DIR / "dm.py"
-DISPATCH_PY = BIN_DIR / "job-dispatch.py"
-
-sys.path.insert(0, str(BIN_DIR))
-try:
-    import pipeline_engine
-    HAS_PIPELINE = True
-except ImportError:
-    HAS_PIPELINE = False
-
-
-def utcnow_dt():
-    return datetime.now(timezone.utc)
-
-
-def utcnow_str():
-    return utcnow_dt().isoformat()
-
-
-def parse_iso(ts_str):
-    if not ts_str:
-        return None
-    try:
-        ts_clean = ts_str.replace("Z", "+00:00")
-        dt = datetime.fromisoformat(ts_clean)
-        if dt.tzinfo is None:
-            dt = dt.replace(tzinfo=timezone.utc)
-        return dt
-    except Exception:
-        return None
-
-
-def load_followups():
-    if not FOLLOWUPS_FILE.exists():
-        return {}
-    try:
-        with open(FOLLOWUPS_FILE, "r", encoding="utf-8") as f:
-            return json.load(f)
-    except Exception:
-        return {}
-
-
-def save_followups(data):
-    tmp_path = f"{FOLLOWUPS_FILE}.tmp.{os.getpid()}"
-    with open(tmp_path, "w", encoding="utf-8") as f:
-        json.dump(data, f, indent=2)
-    os.replace(tmp_path, FOLLOWUPS_FILE)
-
-
-def append_job_log(entry):
-    os.makedirs(os.path.dirname(os.path.abspath(JOB_LOG)), exist_ok=True)
-    with open(JOB_LOG, "a", encoding="utf-8") as f:
-        f.write(json.dumps(entry) + "\n")
-
-
-def send_dm(sender, recipient, target, text):
-    """Dispatch a DM via dm.py. The sweeper's delivery target is explicit
-    followup state (or explicit in-code escalation routing), so pass the
-    main-chat opt-in when the target is main (sidechat-first policy)."""
-    cmd = [
-        sys.executable,
-        str(DM_PY),
-        "send",
-        "--agent", sender,
-        "--to", recipient,
-        "--target", target,
-    ] + (["--allow-main-chat"] if target == "main" else []) + [
-        text,
-    ]
-    try:
-        res = subprocess.run(cmd, capture_output=True, text=True, timeout=90)
-        return res.returncode == 0, res.stdout.strip() or res.stderr.strip()
-    except Exception as e:
-        return False, str(e)
-
-_DM_ID_RE = re.compile(r"\bDM ([0-9a-f]{8})\b")
-_THREAD_UUID_RE = re.compile(
-    r"^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$",
-    re.IGNORECASE,
-)
-
-def _resolve_nudge_thread_uuid(nudge_output):
-    """Parse the DM id from dm.py's stdout, scan the tail (last 5000 lines)
-    of dm-log.jsonl for that id's sidechat_autoprovisioned event, falling
-    back to the sent event's tags.thread. Returns the UUID or None."""
-    m = _DM_ID_RE.search(nudge_output or "")
-    if not m:
-        return None
-    nudge_id = m.group(1)
-    fallback = None
-    try:
-        with open(DM_LOG, "r", encoding="utf-8") as f:
-            lines = f.readlines()
-    except FileNotFoundError:
-        return None
-    for line in lines[-5000:]:
-        line = line.strip()
-        if not line:
-            continue
-        try:
-            e = json.loads(line)
-        except Exception:
-            continue
-        if e.get("id") != nudge_id:
-            continue
-        if e.get("type") == "sidechat_autoprovisioned":
-            uuid = e.get("thread_uuid") or ""
-            if _THREAD_UUID_RE.fullmatch(uuid):
-                return uuid
-        elif e.get("type") == "sent":
-            cand = ((e.get("tags") or {}).get("thread")) or ""
-            if _THREAD_UUID_RE.fullmatch(cand):
-                fallback = cand
-    return fallback
-
-
-_JOB_ID_RE = re.compile(r"^(.+)-(\d{8})-(\d{6})-([0-9a-f]{8})$")
-_OP_NAME_RE = re.compile(r"^[a-z][a-z0-9_.]{0,63}$")
-_JOB_NAME_RE = re.compile(r"^[a-z0-9-]{1,64}$")
-_FALLBACK_RETRY_S = 3600
-
-
-def fallback_due(rec, now=None):
-    """True when a terminal followup should (re)attempt its on_no_result fallback.
-
-    Fires once per record; a failed attempt may retry after _FALLBACK_RETRY_S.
-    Shared by the sweeper terminal path and gravity remediate so the two
-    firing paths can never double-execute.
-    """
-    fb = rec.get("fallback") or {}
-    if fb.get("ran"):
-        return False
-    ts = fb.get("ts")
-    if not ts:
-        return True
-    try:
-        last = datetime.fromisoformat(str(ts).replace("Z", "+00:00"))
-    except Exception:
-        return True
-    if last.tzinfo is None:
-        last = last.replace(tzinfo=timezone.utc)
-    base = now or utcnow_dt()
-    return (base - last).total_seconds() >= _FALLBACK_RETRY_S
-
-_EXEC_OPS_MOD = None
-
-
-def find_job_file(job_name):
-    """Locate a job definition in JOBS_DIR or any archive subdirectory."""
-    p = JOBS_DIR / f"{job_name}.json"
-    if p.exists():
-        return p
-    for match in JOBS_DIR.glob(f"archive/**/{job_name}.json"):
-        if match.is_file():
-            return match
-    for match in JOBS_DIR.glob(f"**/archive/**/{job_name}.json"):
-        if match.is_file():
-            return match
-    return None
-
-
-def derive_job_name(job_id):
-    """Extract the job name from a dispatched job_id (<name>-YYYYMMDD-HHMMSS-<hex8>)."""
-    m = _JOB_ID_RE.match(job_id or "")
-    if not m:
-        return None
-    name = m.group(1)
-    if not find_job_file(name):
-        return None
-    return name
-
-
-def load_job_fallback(job_name):
-    """Return (spec, error) for a job's on_no_result fallback.
-
-    spec is None when the job declares none. Shape:
-      {"job": "<job-name>"}             -> dispatch a fallback job, or
-      {"op": "<exec-op>", "args": {...}} -> run one exec-constrained op.
-    """
-    job_path = find_job_file(job_name)
-    if not job_path:
-        return None, f"unreadable job {job_name}: [Errno 2] No such file or directory: '{JOBS_DIR / (job_name + '.json')}'"
-    try:
-        with open(job_path, "r", encoding="utf-8") as f:
-            cfg = json.load(f)
-    except Exception as e:
-        return None, f"unreadable job {job_name}: {e}"
-    spec = cfg.get("on_no_result")
-    if spec is None:
-        return None, None
-    if not isinstance(spec, dict) or set(spec) - {"job", "op", "args"}:
-        return None, "on_no_result must be an object with job|op (+args)"
-    if bool(spec.get("job")) == bool(spec.get("op")):
-        return None, "on_no_result needs exactly one of job|op"
-    if spec.get("job"):
-        jn = spec["job"]
-        if not isinstance(jn, str) or not _JOB_NAME_RE.fullmatch(jn):
-            return None, "on_no_result.job must be a valid job name"
-        if not find_job_file(jn):
-            return None, f"on_no_result.job {jn!r} does not exist"
-    else:
-        if not isinstance(spec.get("op"), str) or not _OP_NAME_RE.fullmatch(spec["op"]):
-            return None, "on_no_result.op must be a valid op name"
-        if "args" in spec and not isinstance(spec["args"], dict):
-            return None, "on_no_result.args must be an object"
-    return spec, None
-
-
-def _load_exec_ops():
-    global _EXEC_OPS_MOD
-    if _EXEC_OPS_MOD is None:
-        import importlib.util
-        mod_spec = importlib.util.spec_from_file_location(
-            "exec_constrained_sweeper", str(BIN_DIR / "exec-constrained.py"))
-        mod = importlib.util.module_from_spec(mod_spec)
-        mod_spec.loader.exec_module(mod)
-        _EXEC_OPS_MOD = mod
-    return _EXEC_OPS_MOD
-
-
-def run_no_result_fallback(rec, dry_run=False):
-    """Execute a job's on_no_result fallback at terminal followup expiry.
-
-    Returns an outcome dict; never raises (failures are outcome data so
-    one bad spec can't break the sweep).
-    """
-    dm_id = rec.get("dm_id", "?")
-    outcome = {"dm_id": dm_id, "ran": False, "mode": None,
-               "configured": False, "detail": "no job fallback"}
-    try:
-        job_name = derive_job_name(rec.get("job_id"))
-        if not job_name:
-            outcome["detail"] = "no resolvable job_id"
-            return outcome
-        spec, err = load_job_fallback(job_name)
-        if err:
-            outcome.update(configured=True, detail=err)
-            return outcome
-        if spec is None:
-            return outcome
-        outcome["configured"] = True
-        if dry_run:
-            outcome.update(mode="dry_run", detail=json.dumps(spec)[:200])
-            return outcome
-        if spec.get("job"):
-            env = os.environ.copy()
-            env["CHAIN_PREV_JOB_ID"] = rec.get("job_id", "")
-            env["CHAIN_PREV_RESULT"] = (
-                f"TIMEOUT: Agent {rec.get('recipient')} gave no result; "
-                f"on_no_result fallback for job {job_name}")
-            cmd = [sys.executable, str(DISPATCH_PY), spec["job"]]
-            try:
-                p = subprocess.run(cmd, capture_output=True, text=True,
-                                   timeout=180, env=env)
-            except Exception as e:
-                outcome.update(mode="job", detail=f"dispatch exception: {e}")
-                return outcome
-            ok = p.returncode == 0
-            outcome.update(ran=ok, mode="job",
-                           detail=(f"dispatched {spec['job']}" if ok
-                                   else f"dispatch failed: {(p.stderr or p.stdout).strip()[:200]}"))
-        else:
-            mod = _load_exec_ops()
-            op = spec["op"]
-            op_spec = mod.OPS.get(op)
-            if op_spec is None:
-                outcome["detail"] = f"unknown op: {op}"
-                return outcome
-            args = dict(spec.get("args") or {})
-            try:
-                clean = op_spec["validate"](args)
-            except Exception as e:
-                outcome["detail"] = f"op validation failed: {e}"
-                return outcome
-            argv = op_spec["build"](clean)
-            try:
-                p = subprocess.run(argv, capture_output=True, text=True,
-                                   timeout=op_spec.get("timeout", 120))
-            except Exception as e:
-                outcome.update(mode="op", detail=f"op exception: {e}")
-                return outcome
-            ok = p.returncode == 0
-            out = (p.stdout or p.stderr or "").strip()
-            outcome.update(ran=ok, mode="op",
-                           detail=(f"{op} ok: {out[:200]}" if ok
-                                   else f"{op} failed rc={p.returncode}: {out[:200]}"))
-    except Exception as e:
-        outcome["detail"] = f"fallback exception: {e}"
-    return outcome
-
-
-def sweep_cycle(dry_run=False):
-    followups = load_followups()
-    if not followups:
-        return {"status": "ok", "pending": 0, "nudges_sent": 0, "escalations": 0}
-
-    now = utcnow_dt()
-    nudges_count = 0
-    escalations_count = 0
-    fallbacks_count = 0
-    modified = False
-
-    for dm_id, rec in list(followups.items()):
-        if rec.get("status") != "pending":
-            continue
-
-        deadline_dt = parse_iso(rec.get("deadline"))
-        if not deadline_dt or now < deadline_dt:
-            continue
-
-        # Deadline has expired!
-        nudges_sent = rec.get("nudges_sent", 0)
-        nudges_allowed = rec.get("nudges_allowed", 2)
-        sender = rec.get("sender", "opm")
-        recipient = rec.get("recipient")
-        orig_target = rec.get("target", "main")
-        thread_uuid = rec.get("thread_uuid")
-
-        # Ghost-followup fail-fast (2026-10-04, operator-main): a null
-        # thread_uuid means the sidechat was never provisioned, so the
-        # harvester can never match a reply. Nudging is pointless -- flag
-        # for manual triage once instead of burning the nudge budget and
-        # escalating a ghost. Main-chat followups are unaffected (the
-        # harvester matches those by target).
-        if thread_uuid is None and orig_target != "main" and not rec.get("needs_review"):
-            rec["needs_review"] = True
-            rec["status"] = "needs_review"
-            rec["review_reason"] = (
-                "ghost: unresolvable, manual triage "
-                f"(thread_uuid null, target={orig_target}; "
-                "reply can never auto-resolve)"
-            )
-            modified = True
-            append_job_log({
-                "ts": utcnow_str(),
-                "type": "followup_ghost_suppressed",
-                "dm_id": dm_id,
-                "recipient": recipient,
-                "target": orig_target,
-                "reason": "ghost: unresolvable, manual triage",
-            })
-            print(f"Sweeper: suppressing ghost followup {dm_id} "
-                  f"(thread_uuid null, target={orig_target}) -> needs_review",
-                  file=sys.stderr)
-            continue
-
-        if nudges_sent < nudges_allowed:
-            # Deliver next nudge
-            nudge_num = nudges_sent + 1
-            is_final = (nudge_num == nudges_allowed)
-            # Routing: In-thread first, Main on final nudge
-            delivery_target = "main" if is_final else orig_target
-
-            nudge_text = (
-                f"[nudge {nudge_num}/{nudges_allowed}] [ref:{dm_id}] "
-                f"Reminder: awaiting reply to request sent at {rec.get('sent_at', 'earlier')}."
-            )
-            if is_final and orig_target != "main":
-                nudge_text += f" (Origin thread: {orig_target})"
-
-            print(f"Sweeper: Sending nudge {nudge_num}/{nudges_allowed} to {recipient}/{delivery_target}...")
-            if not dry_run:
-                ok, out = send_dm(sender, recipient, delivery_target, nudge_text)
-                if ok:
-                    nudges_count += 1
-                    rec["nudges_sent"] = nudge_num
-                    rec["last_nudge_at"] = utcnow_str()
-                    # Calculate interval for next nudge: proportional to timeout or default 10m
-                    timeout_s = rec.get("timeout_s", 1800)
-                    step_s = max(300, timeout_s // (nudges_allowed + 1))
-                    rec["deadline"] = (now + timedelta(seconds=step_s)).isoformat()
-                    # C1: backfill the thread the nudge actually landed in.
-                    landed_uuid = _resolve_nudge_thread_uuid(out)
-                    if landed_uuid and rec.get("thread_uuid") != landed_uuid:
-                        rec["thread_uuid"] = landed_uuid
-                        print(f"Sweeper: backfilled thread_uuid={landed_uuid} "
-                              f"for followup {dm_id}", file=sys.stderr)
-                    # C2: record final-nudge routing for the harvester.
-                    if is_final:
-                        rec["final_nudge_target"] = "main"
-                    modified = True
-                    append_job_log({
-                        "ts": utcnow_str(),
-                        "type": "followup_nudged",
-                        "dm_id": dm_id,
-                        "nudge_num": nudge_num,
-                        "recipient": recipient,
-                        "target": delivery_target,
-                        "thread_uuid": rec.get("thread_uuid"),
-                    })
-                else:
-                    print(f"Sweeper WARNING: nudge send failed: {out}", file=sys.stderr)
-                    # Failure-mode fix (2026-10-04): advance state on send
-                    # failure so a failing nudge is never re-fired every
-                    # timer tick. failed_sends is tracked separately from
-                    # nudges_sent so a delivery failure does not consume a
-                    # real nudge. Backoff: 5 min base, doubling per failure.
-                    failed = rec.get("failed_sends", 0) + 1
-                    rec["failed_sends"] = failed
-                    rec["last_failure_at"] = utcnow_str()
-                    backoff_s = 300 * (2 ** min(failed - 1, 4))  # 5m,10m,20m,40m,80m cap
-                    rec["deadline"] = (now + timedelta(seconds=backoff_s)).isoformat()
-                    modified = True
-                    append_job_log({
-                        "ts": utcnow_str(),
-                        "type": "followup_nudge_failed",
-                        "dm_id": dm_id,
-                        "nudge_num": nudge_num,
-                        "recipient": recipient,
-                        "target": delivery_target,
-                        "failed_sends": failed,
-                        "backoff_s": backoff_s,
-                        "error": str(out)[:200],
-                    })
-                    # After 3 consecutive failures, stop retrying blindly and
-                    # flag for manual review (delivery may be uncertain or the
-                    # target may be permanently broken).
-                    if failed >= 3:
-                        rec["needs_review"] = True
-                        rec["review_reason"] = (
-                            f"nudge send failed {failed} times consecutively "
-                            f"(last: {str(out)[:120]})"
-                        )
-                        append_job_log({
-                            "ts": utcnow_str(),
-                            "type": "followup_needs_review",
-                            "dm_id": dm_id,
-                            "recipient": recipient,
-                            "failed_sends": failed,
-                        })
-            else:
-                nudges_count += 1
-
-        else:
-            # All nudges exhausted: Terminal escalation
-            escalate_to = rec.get("escalate_to", "opm")
-            esc_text = (
-                f"[ESCALATION] Agent {recipient} failed to reply to DM {dm_id} "
-                f"after {nudges_allowed} nudges. Target was: {orig_target} "
-                f"(thread: {thread_uuid or 'n/a'}). Request sent: {rec.get('sent_at')}."
-            )
-            print(f"Sweeper: Escalating expired follow-up {dm_id} to {escalate_to}...")
-            if not dry_run:
-                ok, out = send_dm("bl", escalate_to, "main", esc_text)
-                rec["status"] = "escalated"
-                rec["escalated_at"] = utcnow_str()
-                escalations_count += 1
-                modified = True
-                append_job_log({
-                    "ts": utcnow_str(),
-                    "type": "followup_escalated",
-                    "dm_id": dm_id,
-                    "recipient": recipient,
-                    "escalated_to": escalate_to,
-                })
-
-                # Check if this dm_id belongs to an active pipeline step
-                if HAS_PIPELINE:
-                    run_entry, step_entry = pipeline_engine.record_step_timeout(dm_id)
-                    if run_entry and step_entry:
-                        j_name = step_entry.get("job_name")
-                        j_file = JOBS_DIR / f"{j_name}.json"
-                        if j_file.exists():
-                            try:
-                                with open(j_file, "r", encoding="utf-8") as f:
-                                    j_cfg = json.load(f)
-                                on_failure = j_cfg.get("on_failure")
-                                if on_failure and (JOBS_DIR / f"{on_failure}.json").exists():
-                                    env = os.environ.copy()
-                                    env["CHAIN_PREV_JOB_ID"] = step_entry.get("job_id")
-                                    env["CHAIN_PREV_RESULT"] = f"TIMEOUT: Agent {recipient} timed out after {nudges_allowed} nudges"
-                                    env["CHAIN_PIPELINE_RUN_ID"] = run_entry.get("run_id")
-                                    next_step_n = step_entry.get("step_n", 1) + 1
-                                    env["CHAIN_STEP_N"] = str(next_step_n)
-                                    cmd = [sys.executable, str(DISPATCH_PY), on_failure,
-                                           "--pipeline-run", run_entry.get("run_id"),
-                                           "--step-n", str(next_step_n)]
-                                    subprocess.Popen(cmd, env=env, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
-                                else:
-                                    pipeline_engine.fail_pipeline(run_entry.get("run_id"), "step_timed_out_without_fallback")
-                            except Exception:
-                                pass
-                # on_no_result fallback: the agent never replied, so run
-                # the job's declared server-side effect now (if any).
-                # fallback_due() dedupes against the gravity firing path.
-                fb = (run_no_result_fallback(rec) if fallback_due(rec)
-                      else {"configured": False, "ran": False, "mode": None,
-                            "detail": "fallback already ran"})
-                if fb["configured"]:
-                    rec["fallback"] = {"ran": fb["ran"], "mode": fb["mode"],
-                                       "detail": fb["detail"][:200],
-                                       "ts": utcnow_str()}
-                    modified = True
-                    append_job_log({
-                        "ts": utcnow_str(),
-                        "type": ("fallback_executed" if fb["ran"]
-                                 else "fallback_failed"),
-                        "dm_id": dm_id,
-                        "recipient": recipient,
-                        "job_id": rec.get("job_id"),
-                        "mode": fb["mode"],
-                        "detail": fb["detail"][:300],
-                    })
-                    print(f"Sweeper: on_no_result fallback for {dm_id}: "
-                          f"ran={fb['ran']} {fb['detail'][:120]}")
-                    if fb["ran"]:
-                        fallbacks_count += 1
-            else:
-                escalations_count += 1
-
-    if modified and not dry_run:
-        save_followups(followups)
-
-    pending_count = sum(1 for r in followups.values() if r.get("status") == "pending")
-    return {
-        "status": "ok",
-        "pending": pending_count,
-        "nudges_sent": nudges_count,
-        "escalations": escalations_count,
-        "fallbacks": fallbacks_count,
-    }
-
-
-def main():
-    parser = argparse.ArgumentParser(description="Autonomous follow-up deadline tracker and sweeper")
-    parser.add_argument("--once", action="store_true", help="Run once and exit (default)")
-    parser.add_argument("--loop", action="store_true", help="Run continuously in a daemon loop")
-    parser.add_argument("--interval", type=int, default=60, help="Interval in seconds for loop (default 60)")
-    parser.add_argument("--dry-run", action="store_true", help="Inspect without sending nudges or updating records")
-    args = parser.parse_args()
-
-    if not args.loop:
-        stats = sweep_cycle(dry_run=args.dry_run)
-        print(f"[{datetime.now(timezone.utc).strftime('%H:%M:%SZ')}] Sweep cycle: {stats['pending']} pending, {stats['nudges_sent']} nudges, {stats['escalations']} escalations.")
-        return
-
-    print(f"Starting follow-up sweeper loop (interval={args.interval}s)...")
-    while True:
-        try:
-            stats = sweep_cycle(dry_run=args.dry_run)
-            print(f"[{datetime.now(timezone.utc).strftime('%H:%M:%SZ')}] Sweep cycle: {stats['pending']} pending, {stats['nudges_sent']} nudges, {stats['escalations']} escalations.")
-        except Exception as e:
-            print(f"ERROR in sweeper loop: {e}", file=sys.stderr)
-        time.sleep(args.interval)
-
-
-if __name__ == "__main__":
-    main()

bin/gravity.py

@@ -1,915 +0,0 @@
-#!/usr/bin/env python3
-"""Timer-driven side-chat adoption: shared gravity library.
-
-Pure Python, no network, no bl dependencies — everything the timers need to
-decide *where* a message goes and *whether* its loop is alive.
-
-Components:
-  GravityConfig      knobs from gravity.json (fail-closed defaults)
-  ThreadRegistry     (recipient, purpose) -> thread_uuid, JSON-backed
-  resolve_target()   explicit target > registry hit > autocreate > fallback
-  LoopState          FIRING/LANDED/SEND_FAILED/SEEN/ANSWERED/NUDGED/
-                     ESCALATED/CLOSED/BROKEN + transition helpers
-  detect_breaks()    loop-break taxonomy detectors over log-derived events
-  actionable_digest() wrap a wake digest as a tracked loop message
-  loop_health()      per-agent ANSWERED+CLOSED / LANDED
-
-The live integrations (job-dispatch.py, dm.py, sidechat-wake.py) import the
-pure functions here; the patches/ directory shows the call-site diffs.
-"""
-
-import json
-import os
-import re
-import sys
-import time
-from pathlib import Path
-
-# ---------------------------------------------------------------- config
-
-DEFAULTS = {
-    "job_default_sidechat": True,
-    "job_sidechat_fallback": "main",
-    "dm_prefer_sidechat_default": False,
-    "dm_sidechat_fallback": "main",
-    "wake_actionable": True,
-    "wake_ack_timeout": 7200,
-    "wake_ack_nudges": 1,
-    "wake_ack_escalate": "opm",
-    "thread_autocreate": True,
-    "loop_silent_ticks": 3,
-    "loop_health_threshold": 0.5,
-}
-
-VALID_AGENTS = ("muse", "pip", "646", "opm")
-UUID_RE = re.compile(
-    r"[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}")
-
-
-class GravityConfig:
-    """Knobs. Unknown keys rejected (typo guard); missing keys get
-    fail-closed defaults (see DEFAULTS)."""
-
-    def __init__(self, data=None):
-        data = data or {}
-        unknown = set(data) - set(DEFAULTS)
-        if unknown:
-            raise ValueError("unknown gravity keys: %s" % sorted(unknown))
-        self._d = dict(DEFAULTS)
-        self._d.update(data)
-
-    @classmethod
-    def load(cls, path):
-        with open(path) as f:
-            return cls(json.load(f))
-
-    def get(self, key):
-        return self._d[key]
-
-    def __getitem__(self, key):
-        return self._d[key]
-
-    def as_dict(self):
-        return dict(self._d)
-
-
-# ---------------------------------------------------------------- registry
-
-class ThreadRegistry:
-    """(recipient, purpose) -> {thread_uuid, created_ts, last_verified_ts}.
-
-    A thread UUID is only valid in the account whose browser created it, so
-    the registry is keyed by recipient — never share UUIDs across accounts.
-
-    Rotation: set()/record_rotation() keep the key pointed at the newest
-    UUID and stash the old one under `<key>:previous` for audit (same
-    model as thread_lifecycle.record_rotation).
-    """
-
-    def __init__(self, path=None, state=None):
-        self.path = path
-        self._s = dict(state or {})
-
-    @classmethod
-    def load(cls, path):
-        try:
-            with open(path) as f:
-                return cls(path=path, state=json.load(f))
-        except (OSError, ValueError):
-            return cls(path=path)
-
-    def save(self):
-        if not self.path:
-            raise ValueError("no path for registry save")
-        tmp = self.path + ".tmp"
-        with open(tmp, "w") as f:
-            json.dump(self._s, f, indent=2)
-        os.replace(tmp, self.path)
-
-    def _key(self, recipient, purpose):
-        return "%s:%s" % (recipient, purpose)
-
-    def get(self, recipient, purpose):
-        """Current thread UUID for (recipient, purpose), following any
-        rotation chain. Returns None if unknown."""
-        return self.resolve(self._key(recipient, purpose))
-
-    def resolve(self, key):
-        """Return the current UUID for a registry key.
-
-        set()/record_rotation() always keep the key pointing at the newest
-        UUID; the `:previous` entry is audit history, not a traversal chain.
-        (Mirrors the guarantee in thread_lifecycle.record_rotation.)
-        """
-        cur = self._s.get(key)
-        if isinstance(cur, str) and UUID_RE.fullmatch(cur):
-            return cur
-        return None
-
-    def previous(self, recipient, purpose):
-        """The rotated-away UUID, if any (for diagnostics)."""
-        return self._s.get(self._key(recipient, purpose) + ":previous")
-
-    def set(self, recipient, purpose, thread_uuid, ts=None):
-        if not UUID_RE.fullmatch(thread_uuid or ""):
-            raise ValueError("not a thread UUID: %r" % thread_uuid)
-        key = self._key(recipient, purpose)
-        old = self._s.get(key)
-        if old and old != thread_uuid:
-            self._s[key + ":previous"] = old
-        self._s[key] = thread_uuid
-        if ts:
-            self._s[key + ":updated_ts"] = ts
-
-    def record_rotation(self, recipient, purpose, new_uuid, ts=None):
-        """Alias for set() — rotation is just a set with history."""
-        self.set(recipient, purpose, new_uuid, ts=ts)
-
-    def known_purposes(self, recipient):
-        prefix = recipient + ":"
-        out = []
-        for k in self._s:
-            if k.startswith(prefix) and ":" not in k[len(prefix):]:
-                out.append(k[len(prefix):])
-        return sorted(out)
-
-
-# ---------------------------------------------------------------- target resolution
-
-def resolve_target(recipient, purpose, cfg, registry,
-                   explicit_target=None, probe=None):
-    """Decide where a timer-driven send goes.
-
-    Order: explicit --target > registry hit (verified) > autocreate >
-    configured fallback. Never raises for routing reasons — worst case is
-    the fallback with a logged reason.
-
-    probe(thread_uuid) -> bool: cheap reachability check (sidechat use).
-    create() -> thread_uuid | None: injected by caller (needs browser).
-
-    Returns (target, thread_uuid_or_None, reason).
-    """
-    if explicit_target:
-        tid = UUID_RE.search(explicit_target)
-        return explicit_target, tid.group(0) if tid else None, "explicit"
-
-    if recipient not in VALID_AGENTS:
-        return cfg["dm_sidechat_fallback"], None, "unknown_recipient"
-
-    uuid = registry.get(recipient, purpose)
-    if uuid:
-        if probe is None or probe(uuid):
-            return uuid, uuid, "registry_hit"
-        return cfg["dm_sidechat_fallback"], None, "registry_stale"
-
-    # No mapping: autocreate or fallback.
-    return None, None, "no_mapping_needs_create"
-
-
-def fallback_target(cfg):
-    """Configured fallback when no sidechat is available. Fails closed
-    toward visibility: main, never drop."""
-    return cfg.get("dm_sidechat_fallback") or "main"
-
-
-# ---------------------------------------------------------------- loop state machine
-
-# Terminal states: CLOSED, ESCALATED, BROKEN, SEND_FAILED.
-STATES = ("FIRING", "LANDED", "SEND_FAILED", "SEEN", "ANSWERED",
-          "NUDGED", "ESCALATED", "CLOSED", "BROKEN")
-
-TERMINAL = frozenset(("SEND_FAILED", "ESCALATED", "CLOSED", "BROKEN"))
-
-# Allowed transitions. NUDGED can cycle (nudge 1..N) until ANSWERED or
-# ESCALATED; BROKEN is reachable from any non-terminal state.
-TRANSITIONS = {
-    "FIRING": ("LANDED", "SEND_FAILED", "BROKEN"),
-    "LANDED": ("SEEN", "ANSWERED", "NUDGED", "BROKEN"),
-    "SEEN": ("ANSWERED", "NUDGED", "BROKEN"),
-    "ANSWERED": ("CLOSED", "BROKEN"),
-    "NUDGED": ("ANSWERED", "NUDGED", "ESCALATED", "BROKEN"),
-    "SEND_FAILED": (),
-    "ESCALATED": (),
-    "CLOSED": (),
-    "BROKEN": (),
-}
-
-
-class LoopState:
-    """One timer-driven loop: timer tick -> tracked follow-up."""
-
-    def __init__(self, loop_id, agent, purpose, thread_uuid=None):
-        self.loop_id = loop_id
-        self.agent = agent
-        self.purpose = purpose
-        self.thread_uuid = thread_uuid
-        self.state = "FIRING"
-        self.nudges = 0
-        self.ticks_without_reply = 0
-        self.history = [("FIRING", None)]
-
-    def transition(self, to, note=None):
-        if to not in TRANSITIONS.get(self.state, ()):
-            raise ValueError("illegal loop transition %s -> %s"
-                             % (self.state, to))
-        self.state = to
-        if to == "NUDGED":
-            self.nudges += 1
-        self.history.append((to, note))
-        return self
-
-    @property
-    def terminal(self):
-        return self.state in TERMINAL
-
-    @property
-    def healthy(self):
-        return self.state in ("ANSWERED", "CLOSED")
-
-    def tick(self):
-        """One scheduler tick with no reply observed."""
-        self.ticks_without_reply += 1
-        return self.ticks_without_reply
-
-
-# ---------------------------------------------------------------- break detectors
-
-def detect_breaks(loop, cfg, checks):
-    """Loop-break taxonomy over caller-supplied check results.
-
-    checks: dict with boolean-ish keys:
-      thread_reachable, signature_ok, scheduler_alive,
-      reply_observed
-    Returns a list of break names (may be empty).
-    """
-    breaks = []
-    if loop.terminal:
-        return breaks
-    if not checks.get("scheduler_alive", True):
-        breaks.append("scheduler_death")
-    if not checks.get("signature_ok", True):
-        breaks.append("auth_rot")
-    if not checks.get("thread_reachable", True):
-        breaks.append("dead_thread")
-    # signature rot: thread was rotated — the stored uuid is stale but a
-    # :previous chain exists. Caller signals via thread_rotated=True and
-    # should already have resolved; flag if not resolved.
-    if checks.get("thread_rotated") and not checks.get("thread_resolved"):
-        breaks.append("signature_rot")
-    if (not checks.get("reply_observed")
-            and loop.ticks_without_reply >= cfg["loop_silent_ticks"]
-            and loop.state in ("LANDED", "SEEN", "NUDGED")):
-        breaks.append("silent_agent")
-    return breaks
-
-
-# ---------------------------------------------------------------- actionable digest
-
-def actionable_digest(digest_text, ack_line=True):
-    """Wrap a wake digest so the timer tick becomes a tracked loop.
-
-    Adds the ack line that the follow-up record's reply closes. The caller
-    sends the result via `dm.py send --expect-reply --thread <uuid>` so a
-    dm_followup record exists — without that send path this is just text.
-    """
-    lines = digest_text.rstrip().split("\n")
-    if ack_line:
-        lines.append("")
-        lines.append("Reply here to acknowledge (closes the loop).")
-    return "\n".join(lines)
-
-
-def dm_send_argv(sender, recipient, target_uuid, message, cfg,
-                 purpose="wake"):
-    """Build the dm.py argv that makes a timer send a *tracked* loop.
-
-    Thread binding (--thread) is what lets nudges route back into the
-    thread instead of falling back to DM.
-    """
-    return [
-        "send",
-        "--agent", sender,
-        "--to", recipient,
-        "--target", target_uuid,
-        "--thread", target_uuid,
-        "--expect-reply",
-        "--reply-timeout", str(cfg["wake_ack_timeout"]),
-        "--reply-nudges", str(cfg["wake_ack_nudges"]),
-        "--reply-escalate", cfg["wake_ack_escalate"],
-        "--tag", "purpose:%s" % purpose,
-        message,
-    ]
-
-
-# ---------------------------------------------------------------- loop health
-
-def loop_health(loops, threshold=0.5):
-    """Per-agent loop health: (ANSWERED + CLOSED) / LANDED.
-
-    loops: iterable of LoopState (or dicts with agent/state keys).
-    Returns {agent: {"landed": n, "answered": n, "health": float|None,
-                     "below_threshold": bool}}.
-    """
-    def _agent(l):
-        return l.agent if isinstance(l, LoopState) else l.get("agent")
-
-    def _state(l):
-        return l.state if isinstance(l, LoopState) else l.get("state")
-
-    # A loop counts as LANDED once it leaves FIRING via LANDED (or beyond).
-    LANDED_OR_BEYOND = ("LANDED", "SEEN", "ANSWERED", "NUDGED",
-                        "ESCALATED", "CLOSED", "BROKEN")
-    acc = {}
-    for l in loops:
-        a, s = _agent(l), _state(l)
-        r = acc.setdefault(a, {"landed": 0, "answered": 0})
-        if s in LANDED_OR_BEYOND:
-            r["landed"] += 1
-        if s in ("ANSWERED", "CLOSED"):
-            r["answered"] += 1
-    out = {}
-    for a, r in acc.items():
-        h = (r["answered"] / r["landed"]) if r["landed"] else None
-        out[a] = {"landed": r["landed"], "answered": r["answered"],
-                  "health": h,
-                  "below_threshold": h is not None and h < threshold}
-    return out
-
-
-# ---------------------------------------------------------------- loop reconstruction & fleet health
-
-NETVM_ROOT = "/home/super/Projects/NetVM"
-FOLLOWUPS_FILE = os.path.join(NETVM_ROOT, "followups.json")
-DM_LOG_FILE = os.path.join(NETVM_ROOT, "dm-log.jsonl")
-JOB_LOG_FILE = os.path.join(NETVM_ROOT, "job-log.jsonl")
-
-
-def reconstruct_loops(limit=50, agent=None, status_filter=None) -> list:
-    """Reconstruct active and recent loops from followups.json and dm-log.jsonl.
-
-    Returns a list of dicts:
-      loop_id, agent, sender, target, purpose, state, sent_at, deadline,
-      nudges_sent, nudges_allowed, escalate_to, tags, summary
-    """
-    loops = {}  # loop_id -> dict
-
-    # 1. Load active/persisted followups
-    if os.path.exists(FOLLOWUPS_FILE):
-        try:
-            with open(FOLLOWUPS_FILE, "r") as f:
-                fdata = json.load(f)
-                if isinstance(fdata, dict):
-                    for did, rec in fdata.items():
-                        recipient = rec.get("recipient")
-                        st = rec.get("status", "pending")
-                        # Map status to loop state
-                        if st == "pending":
-                            lstate = "NUDGED" if rec.get("nudges_sent", 0) > 0 else "LANDED"
-                        elif st == "escalated":
-                            lstate = "ESCALATED"
-                        elif st in ("resolved", "closed"):
-                            lstate = "CLOSED"
-                        else:
-                            lstate = st.upper()
-
-                        loops[did] = {
-                            "loop_id": did,
-                            "agent": recipient,
-                            "sender": rec.get("sender", "super"),
-                            "target": rec.get("target", "main"),
-                            "thread_uuid": rec.get("thread_uuid"),
-                            "purpose": rec.get("route") or "followup",
-                            "state": lstate,
-                            "sent_at": rec.get("sent_at", ""),
-                            "deadline": rec.get("deadline", ""),
-                            "timeout_s": rec.get("timeout_s", 3600),
-                            "nudges_sent": rec.get("nudges_sent", 0),
-                            "nudges_allowed": rec.get("nudges_allowed", 2),
-                            "escalate_to": rec.get("escalate_to", "opm"),
-                            "tags": {"reply:expected": True},
-                            "summary": f"Follow-up for DM {did}",
-                            "source": "followups.json",
-                        }
-        except Exception:
-            pass
-
-    # 2. Extract tracked DMs from dm-log.jsonl
-    answers_map = {}  # id or ref -> answer event
-    dm_events = []
-    failed_ids = set()  # DM ids that failed to send - never delivered
-    if os.path.exists(DM_LOG_FILE):
-        try:
-            with open(DM_LOG_FILE, "r") as f:
-                for line in f:
-                    line = line.strip()
-                    if not line:
-                        continue
-                    try:
-                        entry = json.loads(line)
-                        ev_type = entry.get("type", "")
-                        msg = entry.get("msg") or ""
-
-                        # Check for answers/results/acks
-                        # Note: "verified" means delivered, NOT answered - do not include it here
-                        if "[RESULT" in msg or "[ACK" in msg or "acknowledged" in msg:
-                            sender = entry.get("agent", "")
-                            # Extract referenced DM id if present
-                            m_ref = re.search(r"\[ref:([a-f0-9-]+)\]", msg) or re.search(r"\[(?:ACK|RESULT)\s+([a-f0-9-]+)", msg)
-                            if m_ref:
-                                answers_map[m_ref.group(1)] = entry
-                            if entry.get("id"):
-                                answers_map[entry["id"]] = entry
-
-                        # Track failed sends - these were never delivered
-                        if ev_type in ("send_failed", "failed"):
-                            if entry.get("id"):
-                                failed_ids.add(entry.get("id"))
-
-                        tags = entry.get("tags") or {}
-                        if tags.get("reply:expected") or "[reply:expected]" in msg:
-                            dm_events.append(entry)
-                    except Exception:
-                        continue
-        except Exception:
-            pass
-
-    # Merge dm_events into loops
-    for entry in dm_events:
-        did = entry.get("id") or entry.get("msg_id") or ""
-        if not did:
-            continue
-        if did in loops:
-            continue  # Already have active record from followups.json
-        if did in failed_ids:
-            continue  # Send failed - never delivered, don't count against agent health
-
-        tags = entry.get("tags") or {}
-        recipient = entry.get("to") or entry.get("recipient") or ""
-        sender = entry.get("agent") or entry.get("sender") or "super"
-        target = entry.get("target") or "main"
-        sent_at = entry.get("ts") or entry.get("sent_at") or ""
-        timeout_s = int(tags.get("reply:timeout") or 3600)
-        nudges_max = int(tags.get("reply:nudges") or 2)
-        esc = tags.get("reply:escalate") or "opm"
-        purpose = tags.get("route") or "dm"
-
-        # Determine state
-        if did in answers_map or f"ref:{did}" in answers_map:
-            lstate = "CLOSED"
-        else:
-            # Check age
-            lstate = "LANDED"
-
-        loops[did] = {
-            "loop_id": did,
-            "agent": recipient,
-            "sender": sender,
-            "target": target,
-            "thread_uuid": tags.get("thread"),
-            "purpose": purpose,
-            "state": lstate,
-            "sent_at": sent_at,
-            "deadline": "",
-            "timeout_s": timeout_s,
-            "nudges_sent": 0,
-            "nudges_allowed": nudges_max,
-            "escalate_to": esc,
-            "tags": tags,
-            "summary": entry.get("msg", "")[:80],
-            "source": "dm-log.jsonl",
-        }
-
-    # Filter and sort
-    result = list(loops.values())
-    if agent:
-        result = [l for l in result if l.get("agent") == agent or l.get("sender") == agent]
-    if status_filter:
-        sf = status_filter.lower()
-        if sf == "active" or sf == "pending":
-            result = [l for l in result if l.get("state") in ("FIRING", "LANDED", "SEEN", "NUDGED", "ESCALATED")]
-        elif sf in ("closed", "resolved"):
-            result = [l for l in result if l.get("state") in ("CLOSED", "ANSWERED")]
-        else:
-            result = [l for l in result if l.get("state", "").lower() == sf]
-
-    # Sort newest first
-    result.sort(key=lambda x: x.get("sent_at", ""), reverse=True)
-    return result[:limit]
-
-
-def get_fleet_loop_health(threshold=None) -> dict:
-    """Calculate fleet loop health per agent and overall verdict."""
-    if threshold is None:
-        try:
-            from variables import Variables
-            threshold = Variables().get_float("loop_health_threshold")
-        except Exception:
-            threshold = 0.5
-
-    loops = reconstruct_loops(limit=100)
-    raw = loop_health(loops, threshold=threshold)
-
-    out = {}
-    for ag in VALID_AGENTS:
-        info = raw.get(ag, {"landed": 0, "answered": 0, "health": None, "below_threshold": False})
-        landed = info.get("landed", 0)
-        answered = info.get("answered", 0)
-        h = info.get("health")
-        if landed == 0:
-            status = "IDLE"
-        elif h is not None and h >= threshold:
-            status = "HEALTHY"
-        else:
-            status = "DEGRADED"
-
-        out[ag] = {
-            "agent": ag,
-            "landed": landed,
-            "answered": answered,
-            "health": h,
-            "health_pct": f"{int(h * 100)}%" if h is not None else "-",
-            "threshold": threshold,
-            "below_threshold": info.get("below_threshold", False),
-            "status": status,
-        }
-
-    total_landed = sum(v["landed"] for v in out.values())
-    total_answered = sum(v["answered"] for v in out.values())
-    overall_h = (total_answered / total_landed) if total_landed > 0 else None
-
-    return {
-        "agents": out,
-        "summary": {
-            "total_landed": total_landed,
-            "total_answered": total_answered,
-            "overall_health": overall_h,
-            "overall_health_pct": f"{int(overall_h * 100)}%" if overall_h is not None else "-",
-            "threshold": threshold,
-            "healthy": overall_h is None or overall_h >= threshold,
-        }
-    }
-
-
-def diagnose_breaks() -> list:
-    """Diagnose break taxonomy across intrinsic loops and support services."""
-    import subprocess
-    breaks = []
-
-    # 1. Scheduler / Timers check
-    try:
-        r = subprocess.run(["systemctl", "--user", "is-system-running"],
-                           capture_output=True, text=True, timeout=5)
-        sys_state = r.stdout.strip()
-        if sys_state in ("offline", "stopped"):
-            breaks.append({
-                "type": "scheduler_death",
-                "severity": "CRITICAL",
-                "component": "systemd",
-                "detail": f"Systemd user instance is {sys_state}",
-                "remedy": "Restart systemd user session or start timer jobs manually."
-            })
-    except Exception as e:
-        breaks.append({
-            "type": "scheduler_death",
-            "severity": "WARNING",
-            "component": "systemd",
-            "detail": f"Could not check systemd status: {e}",
-            "remedy": "Verify systemctl --user is available."
-        })
-
-    # 2. SSH key signature check
-    key_path = os.path.expanduser("~/.ssh/id_ed25519")
-    if not os.path.exists(key_path):
-        breaks.append({
-            "type": "auth_rot",
-            "severity": "CRITICAL",
-            "component": "ssh-keys",
-            "detail": f"SSH signing key {key_path} not found",
-            "remedy": "Generate ed25519 key at ~/.ssh/id_ed25519 for cryptographically signed DMs."
-        })
-
-    # 3. Active follow-up loops check
-    active_loops = reconstruct_loops(limit=20, status_filter="pending")
-    now_ts = time.time()
-    for l in active_loops:
-        nudges_sent = l.get("nudges_sent", 0)
-        nudges_max = l.get("nudges_allowed", 2)
-        if nudges_sent >= nudges_max and l.get("state") == "ESCALATED":
-            breaks.append({
-                "type": "silent_agent",
-                "severity": "WARNING",
-                "loop_id": l["loop_id"],
-                "agent": l["agent"],
-                "detail": f"Agent {l['agent']} silent after {nudges_sent}/{nudges_max} nudges for DM {l['loop_id']}",
-                "remedy": f"Check agent {l['agent']} browser tab with 'super fleet status' or nudge via 'super dm send'."
-            })
-
-    # 4. Check for agents held up on approvals
-    try:
-        import approvals
-        fleet_apps = approvals.check_fleet_approvals()
-        for app in fleet_apps:
-            if app.get("has_pending"):
-                node = app["node"]
-                is_trusted = app.get("is_trusted", False)
-                ip = app.get("target") or app.get("ip") or "unknown target"
-                breaks.append({
-                    "type": "approval_blocked",
-                    "severity": "WARNING" if is_trusted else "CRITICAL",
-                    "component": f"node:{node}",
-                    "agent": node,
-                    "detail": f"Agent {node} is held up on browser approval for {ip}",
-                    "remedy": f"Run 'box approvals auto' or 'box approvals allow {node}'."
-                })
-            for w in app.get("input_waits") or []:
-                node = app["node"]
-                breaks.append({
-                    "type": "input_wait",
-                    "severity": "WARNING",
-                    "component": f"node:{node}",
-                    "agent": node,
-                    "detail": f"Agent {node} task '{w.get('task')}' is waiting: {w.get('status')} ({w.get('when')})",
-                    "remedy": f"Open {node}'s task and answer it, or 'box approvals check --node {node}'."
-                })
-    except Exception:
-        pass
-
-    return breaks
-
-
-def _load_sweeper_module():
-    import importlib.util
-    mod_spec = importlib.util.spec_from_file_location(
-        "followup_sweeper_gravity",
-        str(Path(__file__).resolve().parent / "followup-sweeper.py"))
-    mod = importlib.util.module_from_spec(mod_spec)
-    mod_spec.loader.exec_module(mod)
-    return mod
-
-
-def maybe_run_terminal_fallback(rec, now_iso, dry_run=False):
-    """Run a job's on_no_result fallback once at terminal followup expiry.
-
-    Returns an outcome dict, or None when the record has no resolvable
-    fallback. Never raises. The sweeper terminal path shares the
-    fallback_due() guard, so the two firing paths can't double-execute.
-    """
-    try:
-        sw = _load_sweeper_module()
-    except Exception as e:
-        return {"ran": False, "mode": None,
-                "detail": f"sweeper import failed: {e}"}
-    try:
-        if not sw.fallback_due(rec):
-            return None
-        job_name = sw.derive_job_name(rec.get("job_id"))
-        if not job_name:
-            return None
-        spec, err = sw.load_job_fallback(job_name)
-        if err or spec is None:
-            return None
-        if dry_run:
-            return {"ran": False, "mode": "dry_run",
-                    "detail": json.dumps(spec)[:200]}
-        out = sw.run_no_result_fallback(rec)
-        rec["fallback"] = {"ran": out["ran"], "mode": out["mode"],
-                           "detail": out["detail"][:200], "ts": now_iso}
-        try:
-            sw.append_job_log({
-                "ts": now_iso,
-                "type": ("fallback_executed" if out["ran"]
-                         else "fallback_failed"),
-                "dm_id": rec.get("dm_id"),
-                "recipient": rec.get("recipient"),
-                "job_id": rec.get("job_id"),
-                "mode": out["mode"],
-                "detail": out["detail"][:300],
-            })
-        except Exception:
-            pass
-        return out
-    except Exception as e:
-        return {"ran": False, "mode": None,
-                "detail": f"fallback exception: {e}"}
-
-
-def remediate_breaks(dry_run=False) -> dict:
-    """Progressively auto-remediate soft loop breakages while escalating hard breakages.
-
-    Soft breakages (auto-healed):
-      - Pending followups that have received an answer in dm-log.jsonl or chat-history
-        are resolved.
-      - Pending followups with expired deadlines and nudges remaining are re-armed
-        and swept immediately.
-
-    Hard breakages (escalated loudly):
-      - silent_agent (nudges exhausted, no response)
-      - auth_rot (missing SSH signing keys)
-      - scheduler_death (systemd user session offline)
-    """
-    import subprocess
-    from datetime import datetime, timezone
-
-    remediated = []
-    escalated = []
-
-    # 1. Check diagnosed hard breaks first
-    breaks = diagnose_breaks()
-    for b in breaks:
-        if b.get("severity") in ("CRITICAL", "WARNING"):
-            escalated.append(b)
-
-    # 2. Check followups.json for soft break healing
-    f_path = Path("/home/super/Projects/NetVM/followups.json")
-    f_modified = False
-    rearm_sweeper = False
-
-    if f_path.exists():
-        try:
-            with open(f_path, "r") as f:
-                fdata = json.load(f)
-        except Exception:
-            fdata = {}
-
-        now_iso = datetime.now(timezone.utc).isoformat()
-
-        # Build answer map from reconstruct_loops
-        loops = reconstruct_loops(limit=200)
-        answered_dms = {
-            l["loop_id"]: l for l in loops if l.get("state") in ("ANSWERED", "CLOSED")
-        }
-
-        for dm_id, rec in fdata.items():
-            if rec.get("status") == "pending":
-                # Check if it was actually answered in logs
-                if dm_id in answered_dms:
-                    remediated.append({
-                        "action": "auto_resolve_answered",
-                        "loop_id": dm_id,
-                        "agent": rec.get("recipient"),
-                        "detail": f"Follow-up {dm_id} received reply in log but was pending in followups.json. Marked resolved."
-                    })
-                    if not dry_run:
-                        rec["status"] = "resolved"
-                        rec["resolved_at"] = now_iso
-                        rec["resolved_note"] = "auto-healed: reply detected in dm-log"
-                        f_modified = True
-                    continue
-
-                # Check if deadline expired and nudges remaining
-                dl_str = rec.get("deadline", "")
-                nudges_sent = rec.get("nudges_sent", 0)
-                nudges_allowed = rec.get("nudges_allowed", 2)
-
-                is_expired = False
-                if dl_str:
-                    try:
-                        dl_dt = datetime.fromisoformat(dl_str.replace("Z", "+00:00"))
-                        if datetime.now(timezone.utc) > dl_dt:
-                            is_expired = True
-                    except Exception:
-                        pass
-
-                if is_expired and nudges_sent < nudges_allowed:
-                    remediated.append({
-                        "action": "rearm_expired_nudge",
-                        "loop_id": dm_id,
-                        "agent": rec.get("recipient"),
-                        "detail": f"Deadline expired for {dm_id} ({nudges_sent}/{nudges_allowed} nudges). Re-arming immediate sweep."
-                    })
-                    if not dry_run:
-                        rec["deadline"] = now_iso
-                        f_modified = True
-                        rearm_sweeper = True
-
-                # Terminal: nudges exhausted and still no reply. Run the
-                # job's on_no_result fallback (server-side guarantee).
-                if is_expired and nudges_sent >= nudges_allowed:
-                    fb = maybe_run_terminal_fallback(rec, now_iso, dry_run)
-                    if fb is not None:
-                        remediated.append({
-                            "action": "terminal_fallback",
-                            "loop_id": dm_id,
-                            "agent": rec.get("recipient"),
-                            "detail": f"on_no_result ran={fb.get('ran')}: {fb.get('detail', '')[:160]}",
-                        })
-                        if not dry_run and "fallback" in rec:
-                            f_modified = True
-
-        if f_modified and not dry_run:
-            tmp = f"{f_path}.tmp.{os.getpid()}"
-            with open(tmp, "w") as f:
-                json.dump(fdata, f, indent=2)
-            os.replace(tmp, f_path)
-
-        if rearm_sweeper and not dry_run:
-            sweeper_py = Path("/home/super/Projects/NetVM/bin/followup-sweeper.py")
-            if sweeper_py.exists():
-                try:
-                    # A single nudge send takes ~10s median; give the sweep
-                    # room to finish or it dies mid-first-send every time.
-                    subprocess.run([sys.executable, str(sweeper_py), "--once"],
-                                   timeout=300)
-                except Exception:
-                    pass
-
-    # Auto-remediate trusted approval blocks
-    try:
-        import approvals
-        fleet_apps = approvals.check_fleet_approvals()
-        for app in fleet_apps:
-            if app.get("has_pending") and app.get("is_trusted") and app.get("status") != "KEY_APPROVAL":
-                node = app["node"]
-                if not dry_run:
-                    approvals.allow_node_approval(node, always=True, caller="loop-remediate")
-                remediated.append({
-                    "type": "approval_auto_allowed",
-                    "agent": node,
-                    "target": app.get("ip"),
-                    "action": f"Auto-approved trusted browser request on {node} ({app.get('ip')})"
-                })
-    except Exception:
-        pass
-
-    # 3. Alert on hard breakages if any exist
-    if escalated and not dry_run:
-        job_log = Path("/home/super/Projects/NetVM/job-log.jsonl")
-        alert_msg = f"[HARD_BREAK_ALERT] Detected {len(escalated)} unresolvable loop failure(s): " + "; ".join(
-            f"{b.get('type')} ({b.get('severity')}): {b.get('detail')}" for b in escalated[:3]
-        )
-        try:
-            with open(job_log, "a", encoding="utf-8") as jf:
-                jf.write(json.dumps({
-                    "ts": datetime.now(timezone.utc).isoformat(),
-                    "type": "hard_break_alert",
-                    "escalated_count": len(escalated),
-                    "items": escalated,
-                    "summary": alert_msg[:280]
-                }) + "\n")
-        except Exception:
-            pass
-
-        # Dispatch DM alert to opm
-        dm_py = Path("/home/super/Projects/NetVM/bin/dm.py")
-        if dm_py.exists():
-            try:
-                subprocess.run([
-                    sys.executable, str(dm_py), "send",
-                    "--agent", "super",
-                    "--to", "opm",
-                    "--target", "main",
-                    alert_msg[:800]
-                ], timeout=15, capture_output=True)
-            except Exception:
-                pass
-
-    return {
-        "ok": True,
-        "remediated": remediated,
-        "escalated": escalated,
-        "dry_run": dry_run,
-        "count": len(remediated),
-    }
-
-
-def main(argv=None):
-    import argparse
-    ap = argparse.ArgumentParser(description="Loop gravity: reconcile and remediate followup loops")
-    ap.add_argument("--remediate", action="store_true",
-                    help="Run remediate_breaks once (what loop-remediator.timer invokes)")
-    ap.add_argument("--dry-run", action="store_true",
-                    help="Report actions without writing state or sending anything")
-    args = ap.parse_args(argv)
-    if not args.remediate:
-        ap.print_help()
-        return 2
-    result = remediate_breaks(dry_run=args.dry_run)
-    print(json.dumps(result, indent=2))
-    return 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())
-
-
-

bin/hatch_menu/__init__.py

@@ -1,53 +0,0 @@
-"""hatch_menu — modular muse.ai settings-menu navigation + toggles.
-
-One module per part so a site change means patching one file:
-  mouse.py    trusted input primitives (real click, escape, close)
-  dialog.py   settings dialog open / tab nav / rows / back / text
-  controls.py generic radio/switch primitives (verify-then-fallback)
-  toggles.py  toggle registry + sessions over controls and tabs
-  tabs/       one module per settings tab (uniform describe())
-
-invite.py consumes this package; `box chromebox` exposes toggles.
-"""
-
-from hatch_menu.dialog import (
-    TAB_NAMES,
-    TABS,
-    click_row,
-    close_settings,
-    describe_rows,
-    dialog_present,
-    dialog_text,
-    go_back,
-    goto_tab,
-    open_settings,
-)
-from hatch_menu.mouse import MouseError, close, escape, real_click
-from hatch_menu.toggles import (
-    MenuError,
-    describe_tab,
-    get_toggle,
-    list_toggles,
-    set_toggle,
-)
-
-__all__ = [
-    "TAB_NAMES",
-    "TABS",
-    "MenuError",
-    "MouseError",
-    "click_row",
-    "close",
-    "close_settings",
-    "describe_rows",
-    "describe_tab",
-    "dialog_present",
-    "dialog_text",
-    "escape",
-    "get_toggle",
-    "go_back",
-    "goto_tab",
-    "list_toggles",
-    "open_settings",
-    "real_click",
-]

bin/hatch_menu/controls.py

@@ -1,248 +0,0 @@
-"""Generic control primitives (ws-level, no tab knowledge).
-
-Radio/switch list + set with verify-then-fallback: synthetic click,
-verify state flipped, else trusted real click, verify again. Tab
-modules build their flows on these; toggles.py adds addressing.
-"""
-import time
-
-from approvals import cdp_evaluate
-from hatch_menu.mouse import real_click
-
-JS_LIST_RADIOS = """(() => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return null;
-  return Array.from(d.querySelectorAll('input[type="radio"]')).map(r => {
-    let head = '', el = r.parentElement, depth = 0;
-    while (el && el !== d && depth < 6) {
-      const h = el.querySelector('h1,h2,h3,h4');
-      if (h && (h.innerText || '').trim()) {
-        head = h.innerText.trim().slice(0, 60);
-        break;
-      }
-      el = el.parentElement;
-      depth += 1;
-    }
-    const lab = r.closest('label');
-    return {value: r.value, checked: !!r.checked, heading: head,
-      aria: r.getAttribute('aria-label') || '',
-      label: lab ? (lab.innerText || '').trim().slice(0, 80) : ''};
-  });
-})()"""
-
-JS_CLICK_RADIO = """((heading, value) => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return 'NO_DIALOG';
-  const radios = Array.from(d.querySelectorAll('input[type="radio"]'));
-  const headOf = (r) => {
-    let el = r.parentElement, depth = 0;
-    while (el && el !== d && depth < 6) {
-      const h = el.querySelector('h1,h2,h3,h4');
-      if (h && (h.innerText || '').trim())
-        return h.innerText.trim().toLowerCase();
-      el = el.parentElement;
-      depth += 1;
-    }
-    return '';
-  };
-  const t = radios.find(r => headOf(r) === heading.toLowerCase()
-    && r.value === value);
-  if (!t) return 'NO_MATCH';
-  t.click();
-  return 'CLICKED';
-})('%s', '%s')"""
-
-JS_RADIO_RECT = """((heading, value) => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return null;
-  const radios = Array.from(d.querySelectorAll('input[type="radio"]'));
-  const headOf = (r) => {
-    let el = r.parentElement, depth = 0;
-    while (el && el !== d && depth < 6) {
-      const h = el.querySelector('h1,h2,h3,h4');
-      if (h && (h.innerText || '').trim())
-        return h.innerText.trim().toLowerCase();
-      el = el.parentElement;
-      depth += 1;
-    }
-    return '';
-  };
-  const t = radios.find(r => headOf(r) === heading.toLowerCase()
-    && r.value === value);
-  if (!t) return null;
-  const r = t.getBoundingClientRect();
-  return {x: r.x + r.width / 2, y: r.y + r.height / 2};
-})('%s', '%s')"""
-
-JS_CLICK_RADIO_ARIA = """((name) => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return 'NO_DIALOG';
-  const n = name.toLowerCase();
-  const t = Array.from(d.querySelectorAll('input[type="radio"]'))
-    .find(r => (r.getAttribute('aria-label') || '').toLowerCase() === n
-      || (r.value || '').toLowerCase() === n);
-  if (!t) return 'NO_MATCH';
-  t.click();
-  return 'CLICKED';
-})('%s')"""
-
-JS_RADIO_ARIA_RECT = """((name) => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return null;
-  const n = name.toLowerCase();
-  const t = Array.from(d.querySelectorAll('input[type="radio"]'))
-    .find(r => (r.getAttribute('aria-label') || '').toLowerCase() === n
-      || (r.value || '').toLowerCase() === n);
-  if (!t) return null;
-  const r = t.getBoundingClientRect();
-  return {x: r.x + r.width / 2, y: r.y + r.height / 2};
-})('%s')"""
-
-JS_LIST_SWITCHES = """(() => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return null;
-  return Array.from(d.querySelectorAll('[role="switch"]')).map(s => {
-    let el = s.parentElement, label = '', depth = 0;
-    while (el && el !== d && depth < 6) {
-      const t = (el.innerText || '').trim().replace(/\\s+/g, ' ');
-      if (t && t.length < 250) { label = t; break; }
-      el = el.parentElement;
-      depth += 1;
-    }
-    return {label: label.slice(0, 200),
-            aria: s.getAttribute('aria-label') || '',
-            checked: s.getAttribute('aria-checked') === 'true'};
-  });
-})()"""
-
-JS_CLICK_SWITCH = """((label) => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return 'NO_DIALOG';
-  const rowOf = (s) => {
-    let el = s.parentElement, depth = 0;
-    while (el && el !== d && depth < 6) {
-      const t = (el.innerText || '').trim();
-      if (t && t.length < 250) return t.toLowerCase();
-      el = el.parentElement;
-      depth += 1;
-    }
-    return '';
-  };
-  const hits = Array.from(d.querySelectorAll('[role="switch"]'))
-    .filter(s => rowOf(s).includes(label.toLowerCase())
-      || (s.getAttribute('aria-label') || '').toLowerCase()
-        .includes(label.toLowerCase()));
-  if (!hits.length) return 'NO_MATCH';
-  if (hits.length > 1) return 'AMBIGUOUS';
-  hits[0].click();
-  return 'CLICKED';
-})('%s')"""
-
-def _eval(ws, js, timeout=8.0):
-    try:
-        return cdp_evaluate(ws, js, timeout=timeout)
-    except Exception:
-        return None
-
-
-VERIFY_TRIES = 10
-VERIFY_PAUSE = 1.5
-
-
-def _poll(check, tries=VERIFY_TRIES, pause=VERIFY_PAUSE):
-    """Poll a state check until true. Fast exit; tolerates slow commits."""
-    for _ in range(tries):
-        try:
-            if check():
-                return True
-        except Exception:
-            pass
-        time.sleep(pause)
-    return False
-
-
-def list_radios(ws):
-    """All dialog radios with heading/label/value/checked (or [])."""
-    rows = _eval(ws, JS_LIST_RADIOS)
-    return rows if isinstance(rows, list) else []
-
-
-def list_switches(ws):
-    """All dialog switches with row label + checked (or [])."""
-    rows = _eval(ws, JS_LIST_SWITCHES)
-    return rows if isinstance(rows, list) else []
-
-
-def radio_state(ws, heading, value):
-    """Checked state of one heading-grouped radio, None if absent."""
-    for r in list_radios(ws):
-        if (r.get("heading") or "").lower() == heading.lower() \
-                and r.get("value") == value:
-            return bool(r.get("checked"))
-    return None
-
-
-def set_radio_by_heading(ws, heading, value):
-    """Set a heading-grouped radio; poll, else trusted click, poll."""
-    if _eval(ws, JS_CLICK_RADIO % (heading, value)) == "CLICKED" \
-            and _poll(lambda: radio_state(ws, heading, value) is True):
-        return True
-    rect = _eval(ws, JS_RADIO_RECT % (heading, value))
-    if not rect or "x" not in rect:
-        return False
-    try:
-        real_click(ws, rect["x"], rect["y"])
-    except Exception:
-        return False
-    return _poll(lambda: radio_state(ws, heading, value) is True)
-
-
-def radio_aria_state(ws, name):
-    """Checked state of one aria-labeled radio, None if absent."""
-    n = name.lower()
-    for r in list_radios(ws):
-        if (r.get("aria") or "").lower() == n \
-                or (r.get("value") or "").lower() == n:
-            return bool(r.get("checked"))
-    return None
-
-
-def set_radio_by_aria(ws, name):
-    """Set an aria-labeled radio; poll, else trusted click, poll."""
-    if _eval(ws, JS_CLICK_RADIO_ARIA % name) == "CLICKED" \
-            and _poll(lambda: radio_aria_state(ws, name) is True):
-        return True
-    rect = _eval(ws, JS_RADIO_ARIA_RECT % name)
-    if not rect or "x" not in rect:
-        return False
-    try:
-        real_click(ws, rect["x"], rect["y"])
-    except Exception:
-        return False
-    return _poll(lambda: radio_aria_state(ws, name) is True)
-
-
-def switch_state(ws, label):
-    """Checked state of one label-matched switch, None if not unique."""
-    matches = [s for s in list_switches(ws)
-               if label.lower() in (s.get("label") or "").lower()
-               or label.lower() in (s.get("aria") or "").lower()]
-    if len(matches) != 1:
-        return None
-    return bool(matches[0].get("checked"))
-
-
-def set_switch(ws, label, on):
-    """Set a switch by row-label/aria match; no-op when already there.
-
-    Single click only: a fallback re-click would UNDO a slow commit
-    (switches toggle). The fresh-session readback decides.
-    """
-    state = switch_state(ws, label)
-    if state is None:
-        return False
-    if state == bool(on):
-        return True
-    if _eval(ws, JS_CLICK_SWITCH % label) != "CLICKED":
-        return False
-    return _poll(lambda: switch_state(ws, label) is bool(on))

bin/hatch_menu/dialog.py

@@ -1,162 +0,0 @@
-"""Settings dialog navigation: open, tabs, rows, back, text.
-
-Open is retried (single-shot opens flake ~1/4 live): each attempt
-re-Escapes and re-drives the dock menu from scratch. Tab clicks are
-idempotent (re-clicking the active tab is a harmless no-op), so goto
-always clicks and reports the click result instead of guessing which
-tab is active.
-"""
-import json
-import time
-
-from approvals import cdp_evaluate
-from hatch_menu.mouse import escape, real_click
-
-TAB_NAMES = ["General", "Connectors", "Wallet", "Secure store",
-             "Permissions", "Messaging channels", "Devices",
-             "Data controls", "Help & support", "Legal info"]
-TABS = TAB_NAMES  # legacy alias
-
-# Tab rail buttons carry bare tab names and live outside any nav
-# landmark, so row matchers exclude them by exact text (live 2026-10-06).
-_JS_TABS = json.dumps(TAB_NAMES)
-
-DOCK_MORE_TESTID = "hatch-dock-more"
-
-JS_DOCK_RECT = ("(() => { const b = document.querySelector("
-                "'[data-testid=\"hatch-dock-more\"]'); if (!b) return null;"
-                " const r = b.getBoundingClientRect();"
-                " return {x: r.x + r.width/2, y: r.y + r.height/2}; })()")
-
-JS_CLICK_SETTINGS_ITEM = ("(() => { const it = Array.from(document."
-    "querySelectorAll('[role=\"menuitem\"]')).find(el => el.getAttribute("
-    "'data-pel-click') === 'settings_nav_click');"
-    " if (!it) return 'NO_ITEM'; it.click(); return 'CLICKED'; })()")
-
-JS_DIALOG_PRESENT = ("(() => !!document.querySelector('[role=\"dialog\"]'))()")
-
-JS_DIALOG_TEXT = ("(() => { const d = document.querySelector("
-                  "'[role=\"dialog\"]'); return d ? d.innerText : null; })()")
-
-JS_GOTO_TAB_TMPL = ("(() => { const b = Array.from(document."
-    "querySelectorAll('[role=\"dialog\"] button')).find(x => "
-    "(x.innerText||'').trim() === '%s');"
-    " if (!b) return 'NO_TAB'; b.click(); return 'CLICKED'; })()")
-
-JS_CLICK_ROW_TMPL = ("((name) => {"
-    " const d = document.querySelector('[role=\"dialog\"]');"
-    " if (!d) return 'NO_DIALOG';"
-    " const TABS = " + _JS_TABS + ";"
-    " const inNav = (el) => !!el.closest("
-    "'nav, [role=\"tablist\"], [role=\"navigation\"]');"
-    " const els = Array.from(d.querySelectorAll("
-    "'button, [role=\"button\"], a')).filter(e => !inNav(e));"
-    " const t = els.find(e => {"
-    " const txt = (e.innerText || '').trim();"
-    " return !TABS.includes(txt) && txt.toLowerCase()"
-    ".startsWith(name.toLowerCase()); });"
-    " if (!t) return 'NO_ROW'; t.click(); return 'CLICKED'; })('%s')")
-
-JS_DESCRIBE_ROWS = ("(() => {"
-    " const d = document.querySelector('[role=\"dialog\"]');"
-    " if (!d) return null;"
-    " const TABS = " + _JS_TABS + ";"
-    " const inNav = (el) => !!el.closest("
-    "'nav, [role=\"tablist\"], [role=\"navigation\"]');"
-    " return Array.from(d.querySelectorAll("
-    "'button, [role=\"button\"], a')).filter(e => !inNav(e))"
-    ".filter(e => !TABS.includes((e.innerText || '').trim()))"
-    ".map(e => { const lines = (e.innerText || '').trim().split('\\n');"
-    " return {name: (lines[0] || '').slice(0, 80),"
-    " detail: lines.slice(1).join(' / ').slice(0, 120)}; }); })()")
-
-JS_GO_BACK = ("(() => { const d = document.querySelector("
-              "'[role=\"dialog\"]'); if (!d) return 'NO_DIALOG';"
-              " const b = Array.from(d.querySelectorAll('button')).find("
-              "x => (x.getAttribute('aria-label') || '') === 'Go back');"
-              " if (!b) return 'NO_BACK'; b.click(); return 'CLICKED'; })()")
-
-
-def _eval(ws, js, timeout=8.0):
-    try:
-        return cdp_evaluate(ws, js, timeout=timeout)
-    except Exception:
-        return None
-
-
-def dialog_present(ws):
-    """True when a dialog is currently open."""
-    return bool(_eval(ws, JS_DIALOG_PRESENT, timeout=5.0))
-
-
-def dialog_text(ws, timeout=8.0, limit=4000):
-    """Inner text of the open dialog, or None."""
-    try:
-        text = cdp_evaluate(ws, JS_DIALOG_TEXT, timeout=timeout)
-    except Exception:
-        return None
-    if not isinstance(text, str) or not text:
-        return None
-    return text[:limit]
-
-
-def open_settings(ws, tries=3):
-    """Open the Settings dialog via the dock menu. True when open."""
-    for _ in range(tries):
-        escape(ws)
-        time.sleep(0.5)
-        rect = _eval(ws, JS_DOCK_RECT, timeout=5.0)
-        if not rect or "x" not in rect:
-            continue
-        try:
-            real_click(ws, rect["x"], rect["y"])
-        except Exception:
-            continue
-        time.sleep(1.0)
-        if _eval(ws, JS_CLICK_SETTINGS_ITEM,
-                 timeout=5.0) != "CLICKED":
-            continue
-        time.sleep(1.5)
-        if dialog_present(ws):
-            return True
-    return False
-
-
-def goto_tab(ws, name, timeout=5.0):
-    """Click a Settings tab by visible name. True when clicked."""
-    if _eval(ws, JS_GOTO_TAB_TMPL % name, timeout=timeout) != "CLICKED":
-        return False
-    time.sleep(0.8)
-    return True
-
-
-def click_row(ws, name, tab=None, timeout=8.0):
-    """Click a content row; verify the drill/expand opened. Bool."""
-    if tab is not None and not goto_tab(ws, tab, timeout=timeout):
-        return False
-    if _eval(ws, JS_CLICK_ROW_TMPL % name, timeout=timeout) != "CLICKED":
-        return False
-    time.sleep(1.0)
-    text = dialog_text(ws, timeout=timeout) or ""
-    return name.lower() in text.lower()
-
-
-def describe_rows(ws, tab=None, timeout=8.0):
-    """Inventory rows (name/detail) on a tab. [] when unreadable."""
-    if tab is not None and not goto_tab(ws, tab, timeout=timeout):
-        return []
-    rows = _eval(ws, JS_DESCRIBE_ROWS, timeout=timeout)
-    return rows if isinstance(rows, list) else []
-
-
-def go_back(ws):
-    """Click the sub-page Go back button. True when clicked."""
-    if _eval(ws, JS_GO_BACK, timeout=5.0) != "CLICKED":
-        return False
-    time.sleep(0.8)
-    return True
-
-
-def close_settings(ws):
-    """Dismiss settings/popovers. Never raises."""
-    escape(ws)

bin/hatch_menu/mouse.py

@@ -1,62 +0,0 @@
-"""Trusted input primitives for menu automation.
-
-Radix triggers (dock menu, permission-mode choosers) ignore synthetic
-JS clicks: they need real CDP Input.dispatchMouseEvent press+release.
-"""
-import json
-import time
-
-from approvals import _cdp_req_ids as _shared_cdp_ids, cdp_evaluate
-
-ESCAPE_JS = ("(() => { document.dispatchEvent(new KeyboardEvent("
-             "'keydown', {key: 'Escape', code: 'Escape',"
-             " bubbles: true})); return 'ESC'; })()")
-
-
-class MouseError(RuntimeError):
-    """Trusted click failed (CDP transport or echo timeout)."""
-
-
-def escape(ws):
-    """Dismiss topmost popover/menu/dialog. Never raises."""
-    try:
-        cdp_evaluate(ws, ESCAPE_JS, timeout=3.0)
-    except Exception:
-        pass
-
-
-def close(ws):
-    """Close a CDP websocket. Never raises."""
-    try:
-        ws.close()
-    except Exception:
-        pass
-
-
-def real_click(ws, x, y, timeout=5.0):
-    """Trusted press+release at page coordinates.
-
-    Shares approvals' monotonic CDP id counter so ids stay unique on
-    the connection; matches responses by id like cdp_evaluate.
-    Raises MouseError on transport or echo-timeout failure.
-    """
-    try:
-        for typ in ("mousePressed", "mouseReleased"):
-            req_id = next(_shared_cdp_ids)
-            ws.send(json.dumps({"id": req_id,
-                                "method": "Input.dispatchMouseEvent",
-                                "params": {"type": typ, "x": x, "y": y,
-                                          "button": "left",
-                                          "clickCount": 1}}))
-            deadline = time.time() + timeout
-            while time.time() < deadline:
-                resp = json.loads(ws.recv())
-                if resp.get("id") == req_id:
-                    break
-            else:
-                raise MouseError("mouse echo timeout for %s" % typ)
-    except MouseError:
-        raise
-    except Exception as e:
-        raise MouseError("real click failed: %s: %s"
-                         % (type(e).__name__, e))

bin/hatch_menu/tabs/__init__.py

@@ -1,30 +0,0 @@
-"""One module per Settings tab. Uniform: TAB, describe(ws)."""
-from hatch_menu.tabs import (
-    connectors,
-    data_controls,
-    devices,
-    general,
-    help_support,
-    legal,
-    messaging,
-    permissions,
-    secure_store,
-    wallet,
-)
-
-TAB_MODULES = {
-    "General": general,
-    "Connectors": connectors,
-    "Wallet": wallet,
-    "Secure store": secure_store,
-    "Permissions": permissions,
-    "Messaging channels": messaging,
-    "Devices": devices,
-    "Data controls": data_controls,
-    "Help & support": help_support,
-    "Legal info": legal,
-}
-
-__all__ = ["TAB_MODULES", "connectors", "data_controls", "devices",
-           "general", "help_support", "legal", "messaging",
-           "permissions", "secure_store", "wallet"]

bin/hatch_menu/tabs/connectors.py

@@ -1,10 +0,0 @@
-"""Connectors tab: read-only inventory (search + per-app Connect/View)."""
-from hatch_menu import dialog
-
-TAB = "Connectors"
-
-
-def describe(ws):
-    """Row inventory + text excerpt (describe-only for now)."""
-    return {"rows": dialog.describe_rows(ws, TAB),
-            "text": (dialog.dialog_text(ws) or "")[:400]}

bin/hatch_menu/tabs/data_controls.py

@@ -1,101 +0,0 @@
-"""Data controls tab: model-improvement switch (read-only otherwise).
-
-The switch label is pinned from live recon; resolution prefers it
-and falls back to single-switch, then keyword match. Sets use a
-trusted click: synthetic clicks are proven no-ops here (2026-10-06).
-Import/Delete rows are inventoried, never touched.
-"""
-import time
-
-from approvals import cdp_evaluate
-from hatch_menu import controls, dialog
-from hatch_menu.mouse import real_click
-
-TAB = "Data controls"
-SWITCH_LABEL = "Help improve our AI models"
-_KEYWORDS = ("improv", "train", "model", "data", "usage")
-
-JS_AI_RECT_TMPL = """((label) => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return null;
-  const rowOf = (s) => {
-    let el = s.parentElement, depth = 0;
-    while (el && el !== d && depth < 6) {
-      const t = (el.innerText || '').trim();
-      if (t && t.length < 250) return t.toLowerCase();
-      el = el.parentElement;
-      depth += 1;
-    }
-    return '';
-  };
-  const hits = Array.from(d.querySelectorAll('[role="switch"]'))
-    .filter(s => rowOf(s).includes(label.toLowerCase())
-      || (s.getAttribute('aria-label') || '').toLowerCase()
-        .includes(label.toLowerCase()));
-  if (hits.length !== 1) return null;
-  const r = hits[0].getBoundingClientRect();
-  return {x: r.x + r.width / 2, y: r.y + r.height / 2};
-})('%s')"""
-
-
-def _eval(ws, js, timeout=8.0):
-    try:
-        return cdp_evaluate(ws, js, timeout=timeout)
-    except Exception:
-        return None
-
-
-def _resolve(ws):
-    """The improvement switch dict, or None when not resolvable."""
-    if not dialog.goto_tab(ws, TAB):
-        return None
-    switches = controls.list_switches(ws)
-    for s in switches:
-        blob = ((s.get("label") or "") + " "
-                + (s.get("aria") or "")).lower()
-        if SWITCH_LABEL.lower() in blob:
-            return s
-    if len(switches) == 1:
-        return switches[0]
-    for kw in _KEYWORDS:
-        for s in switches:
-            blob = ((s.get("label") or "") + " "
-                    + (s.get("aria") or "")).lower()
-            if kw in blob:
-                return s
-    return None
-
-
-def ai_improvement(ws):
-    """Improvement-switch state: True/False, None when unreadable."""
-    sw = _resolve(ws)
-    return None if sw is None else bool(sw.get("checked"))
-
-
-def set_ai_improvement(ws, on):
-    """Set via one trusted click; synthetic clicks are no-ops. Bool."""
-    sw = _resolve(ws)
-    if sw is None:
-        return False
-    if bool(sw.get("checked")) == bool(on):
-        return True
-    rect = _eval(ws, JS_AI_RECT_TMPL % SWITCH_LABEL)
-    if not rect or "x" not in rect:
-        return False
-    try:
-        real_click(ws, rect["x"], rect["y"])
-    except Exception:
-        return False
-    for _ in range(8):
-        time.sleep(2.0)
-        if ai_improvement(ws) is bool(on):
-            return True
-    return False
-
-
-def describe(ws):
-    """Inventory: switch state + row names (import/delete read-only)."""
-    state = ai_improvement(ws)
-    return {"ai_improvement":
-            ("on" if state else "off") if state is not None else None,
-            "rows": dialog.describe_rows(ws, TAB)}

bin/hatch_menu/tabs/devices.py

@@ -1,13 +0,0 @@
-"""Devices tab: read-only inventory (paired devices or empty state)."""
-from hatch_menu import dialog
-
-TAB = "Devices"
-
-
-def describe(ws):
-    """Row inventory + empty flag (describe-only for now)."""
-    text = dialog.dialog_text(ws) or ""
-    low = text.lower()
-    return {"rows": dialog.describe_rows(ws, TAB),
-            "empty": ("no devices" in low or "don't have" in low),
-            "text": text[:400]}

bin/hatch_menu/tabs/general.py

@@ -1,95 +0,0 @@
-"""General tab: usage balances, theme picker, redeem entrypoint.
-
-Usage parsing moved here from invite.py (single copy). All flows are
-ws-level: sessions and error shaping live in toggles.py.
-"""
-import re
-
-from approvals import cdp_evaluate
-from hatch_menu import controls, dialog
-
-TAB = "General"
-
-THEME_VALUES = ("avatar", "default", "blue", "purple", "pink",
-                "orange", "green", "beige", "monochrome")
-
-_FREE_RE = re.compile(r"\bfree plan\b", re.IGNORECASE)
-_PCT_RE = re.compile(r"(\d+)%\s*used")
-_RESET_RE = re.compile(r"resets?\s+on\s+([A-Z][a-z]+\s+\d{1,2})",
-                       re.IGNORECASE)
-_TOK_RE = re.compile(r"\(([0-9.,]+\s*[BMK]?)\s*tokens?\s+left\)",
-                     re.IGNORECASE)
-
-
-def parse_usage_text(text):
-    """Parse a General-tab usage block into a balance dict.
-
-    Returns None for empty/unreadable text. Weekly fields stay None
-    when the block only carries additional-tokens rows.
-    """
-    if not text or not text.strip():
-        return None
-    out = {"plan": None, "weekly": {"pct_used": None, "resets_on": None},
-           "additional": {"pct_used": None, "tokens_left": None,
-                          "never_expires": False}}
-    if _FREE_RE.search(text):
-        out["plan"] = "free"
-    pcts = _PCT_RE.findall(text)
-    if pcts:
-        out["weekly"]["pct_used"] = int(pcts[0])
-        if len(pcts) > 1:
-            out["additional"]["pct_used"] = int(pcts[1])
-    m = _RESET_RE.search(text)
-    if m:
-        out["weekly"]["resets_on"] = m.group(1)
-    m = _TOK_RE.search(text)
-    if m:
-        out["additional"]["tokens_left"] = m.group(1).strip()
-    if "never expires" in text.lower():
-        out["additional"]["never_expires"] = True
-    return out
-
-
-def _eval(ws, js, timeout=8.0):
-    try:
-        return cdp_evaluate(ws, js, timeout=timeout)
-    except Exception:
-        return None
-
-
-def usage(ws):
-    """Usage balances from General tab. Dict, or None when unreadable."""
-    if not dialog.goto_tab(ws, TAB):
-        return None
-    return parse_usage_text(dialog.dialog_text(ws))
-
-
-def theme(ws):
-    """Current theme radio value, or None when unreadable."""
-    if not dialog.goto_tab(ws, TAB):
-        return None
-    for r in controls.list_radios(ws):
-        if r.get("checked"):
-            return r.get("value")
-    return None
-
-
-def set_theme(ws, value):
-    """Set theme by aria-label; verify-then-fallback. Bool."""
-    if not dialog.goto_tab(ws, TAB):
-        return False
-    return controls.set_radio_by_aria(ws, value)
-
-
-def redeem_row(ws):
-    """True when the 'Redeem invite code' row is present (eligible)."""
-    if not dialog.goto_tab(ws, TAB):
-        return False
-    text = (dialog.dialog_text(ws) or "").lower()
-    return "redeem" in text and "invite code" in text
-
-
-def describe(ws):
-    """Full General inventory: usage, theme, redeem entrypoint."""
-    return {"usage": usage(ws), "theme": theme(ws),
-            "redeem_row": redeem_row(ws)}

bin/hatch_menu/tabs/help_support.py

@@ -1,10 +0,0 @@
-"""Help & support tab: read-only inventory (help links)."""
-from hatch_menu import dialog
-
-TAB = "Help & support"
-
-
-def describe(ws):
-    """Row inventory + text excerpt (describe-only for now)."""
-    return {"rows": dialog.describe_rows(ws, TAB),
-            "text": (dialog.dialog_text(ws) or "")[:400]}

bin/hatch_menu/tabs/legal.py

@@ -1,10 +0,0 @@
-"""Legal info tab: read-only inventory (legal links)."""
-from hatch_menu import dialog
-
-TAB = "Legal info"
-
-
-def describe(ws):
-    """Row inventory + text excerpt (describe-only for now)."""
-    return {"rows": dialog.describe_rows(ws, TAB),
-            "text": (dialog.dialog_text(ws) or "")[:400]}

bin/hatch_menu/tabs/messaging.py

@@ -1,10 +0,0 @@
-"""Messaging channels tab: read-only inventory (channel rows)."""
-from hatch_menu import dialog
-
-TAB = "Messaging channels"
-
-
-def describe(ws):
-    """Row inventory + text excerpt (describe-only for now)."""
-    return {"rows": dialog.describe_rows(ws, TAB),
-            "text": (dialog.dialog_text(ws) or "")[:400]}

bin/hatch_menu/tabs/permissions.py

@@ -1,422 +0,0 @@
-"""Permissions tab: defaults radios, website modes, protocols, advanced.
-
-Contracts live here (single copy): toggles.py references these
-constants for addressing. Drills open sub-pages, read, and come back
-via the back affordance with tab re-entry fallback. All flows are
-ws-level; sessions and error shaping live in toggles.py.
-"""
-import re
-import time
-
-from approvals import cdp_evaluate
-from hatch_menu import controls, dialog
-from hatch_menu.mouse import escape, real_click
-
-TAB = "Permissions"
-ROOT_MARK = "Manage permissions"
-
-CONNECTOR_HEADING = "Connector defaults"
-WEB_HEADING = "Web access defaults"
-DEFAULT_VALUES = ("auto_allow", "always_ask")
-
-WEBSITE_MODES = ("Allow", "Ask", "Deny")
-
-ADV_LABELS = {"transparent_proxy": "Transparent proxy",
-              "tls_interception": "TLS interception",
-              "sni_mismatch_rejection": "SNI mismatch rejection"}
-
-# Row titles (first line of each protocol row) pinned live 2026-10-06:
-# network primitives on every node checked; MCP titles kept
-# defensively in case they appear on other plans/accounts.
-PROTOCOL_SLUGS = {"Outbound SSH": "outbound-ssh",
-                  "Outgoing email (SMTP)": "smtp",
-                  "Email mailbox access (IMAP, POP3)": "imap-pop3",
-                  "Database connections": "database",
-                  "File transfer (FTP)": "ftp",
-                  "External DNS lookups": "dns",
-                  "Other TCP connections": "other-tcp",
-                  "Other UDP traffic": "other-udp",
-                  "Model Context Protocol servers (SSE)": "mcp-sse",
-                  "Model Context Protocol servers (Streamable HTTP)":
-                  "mcp-streamable",
-                  "Agent Skills endpoints": "agent-skills",
-                  "MCP Apps (UI extensions)": "mcp-apps",
-                  "MCP remote OAuth": "mcp-oauth"}
-
-JS_WEBSITES = """(() => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return null;
-  const out = [];
-  for (const b of d.querySelectorAll('button')) {
-    const m = (b.getAttribute('aria-label') || '').match(
-      /^Change permission mode for (.+),\\s*(Allow|Ask|Deny)$/i);
-    if (!m) continue;
-    const r = b.getBoundingClientRect();
-    out.push({host: m[1].trim(), mode: m[2],
-      x: r.x + r.width / 2, y: r.y + r.height / 2});
-  }
-  return out;
-})()"""
-
-JS_MODE_MENU = """(() => {
-  return Array.from(document.querySelectorAll('[role="menuitem"]'))
-    .map(m => ({text: (m.innerText || '').trim()}));
-})()"""
-
-JS_CLICK_MODE = """((mode) => {
-  const m = Array.from(document.querySelectorAll('[role="menuitem"]'))
-    .find(el => (el.innerText || '').trim() === mode);
-  if (!m) return 'NO_MATCH';
-  m.click();
-  return 'CLICKED';
-})('%s')"""
-
-JS_PROTO_ROWS = """(() => {
-  const d = document.querySelector('[role="dialog"]');
-  if (!d) return null;
-  return Array.from(d.querySelectorAll('[role="switch"]')).map(s => {
-    let el = s.parentElement, title = '', depth = 0;
-    while (el && el !== d && depth < 6) {
-      const t = (el.innerText || '').trim().split('\\n')[0] || '';
-      if (t) { title = t.slice(0, 80); break; }
-      el = el.parentElement;
-      depth += 1;
-    }
-    const r = s.getBoundingClientRect();
-    return {title: title,
-            checked: s.getAttribute('aria-checked') === 'true',
-            x: r.x + r.width / 2, y: r.y + r.height / 2};
-  });
-})()"""
-
-
-def _eval(ws, js, timeout=8.0):
-    try:
-        return cdp_evaluate(ws, js, timeout=timeout)
-    except Exception:
-        return None
-
-
-def _stable_rows(ws, js, retries=4, pause=1.5):
-    """Repeat a row read until two consecutive reads agree.
-
-    Guards mid-animation partial DOM (innerText shifts while the
-    sub-page slides in). Returns the agreed list, or None.
-    """
-    last = "sentinel"
-    for _ in range(retries):
-        rows = _eval(ws, js)
-        if isinstance(rows, list) and rows == last:
-            return rows
-        last = rows if isinstance(rows, list) else "sentinel"
-        time.sleep(pause)
-    return last if isinstance(last, list) else None
-
-
-def _slug(title):
-    """Protocol slug: registry hit, else slugified, else None."""
-    if title in PROTOCOL_SLUGS:
-        return PROTOCOL_SLUGS[title]
-    clean = re.sub(r"[^a-z0-9]+", "-",
-                   title.strip().lower()).strip("-")
-    return clean or None
-
-
-def _canon_mode(mode):
-    """Canonical Allow/Ask/Deny (case-insensitive); passthrough else."""
-    for m in WEBSITE_MODES:
-        if (mode or "").lower() == m.lower():
-            return m
-    return mode
-
-
-def resolve_protocol(name):
-    """Slug/title -> row title, None when unresolvable.
-
-    Exact slug or title first; then a unique case-insensitive
-    substring over titles+slugs (so 'ssh' finds Outbound SSH).
-    """
-    if not isinstance(name, str) or not name.strip():
-        return None
-    want = name.strip().lower()
-    for title, slug in PROTOCOL_SLUGS.items():
-        if want == slug or want == title.lower():
-            return title
-    hits = [t for t, s in PROTOCOL_SLUGS.items()
-            if want in t.lower() or want in s]
-    if len(hits) == 1:
-        return hits[0]
-    return None
-
-
-def _back_to_root(ws):
-    """Back affordance, else tab re-entry; verify root text."""
-    dialog.go_back(ws)
-    if ROOT_MARK in (dialog.dialog_text(ws) or ""):
-        return True
-    if not dialog.goto_tab(ws, TAB):
-        return False
-    return ROOT_MARK in (dialog.dialog_text(ws) or "")
-
-
-def defaults(ws):
-    """Connector + web default values (each value or None)."""
-    if not dialog.goto_tab(ws, TAB):
-        return {"connector_defaults": None, "web_access": None}
-    heads = {CONNECTOR_HEADING.lower(): "connector_defaults",
-             WEB_HEADING.lower(): "web_access"}
-    vals = {CONNECTOR_HEADING.lower(): [], WEB_HEADING.lower(): []}
-    for r in controls.list_radios(ws):
-        h = (r.get("heading") or "").lower()
-        if h in vals and r.get("checked"):
-            vals[h].append(r.get("value"))
-    out = {}
-    for h, key in heads.items():
-        out[key] = vals[h][0] if len(vals[h]) == 1 else None
-    return out
-
-
-def set_default(ws, which, value):
-    """Set one defaults radio. Bool."""
-    if not dialog.goto_tab(ws, TAB):
-        return False
-    heading = CONNECTOR_HEADING if which == "connector_defaults" \
-        else WEB_HEADING
-    return controls.set_radio_by_heading(ws, heading, value)
-
-
-def ensure_advanced(ws):
-    """Expand Advanced network settings when collapsed. Bool."""
-    if not dialog.goto_tab(ws, TAB):
-        return False
-    labels = [s.get("label", "") for s in controls.list_switches(ws)]
-    if any("Transparent proxy" in lab for lab in labels):
-        return True
-    if not dialog.click_row(ws, "Advanced network settings", TAB):
-        return False
-    time.sleep(0.6)
-    labels = [s.get("label", "") for s in controls.list_switches(ws)]
-    return any("Transparent proxy" in lab for lab in labels)
-
-
-def advanced(ws):
-    """Advanced switch states {key: on/off/None}."""
-    if not ensure_advanced(ws):
-        return {k: None for k in ADV_LABELS}
-    out = {}
-    for key, label in ADV_LABELS.items():
-        state = None
-        for s in controls.list_switches(ws):
-            if label.lower() in (s.get("label") or "").lower():
-                state = bool(s.get("checked"))
-                break
-        out[key] = ("on" if state else "off") if state is not None \
-            else None
-    return out
-
-
-def set_advanced(ws, key, on):
-    """Set one advanced switch. Bool."""
-    if key not in ADV_LABELS or not ensure_advanced(ws):
-        return False
-    return controls.set_switch(ws, ADV_LABELS[key], on)
-
-
-def _websites_raw(ws):
-    """Drill into Websites; rows or None (stays on sub-page)."""
-    if not dialog.click_row(ws, "Websites", TAB):
-        return None
-    return _stable_rows(ws, JS_WEBSITES)
-
-
-def websites(ws):
-    """[{host, mode}] (back at root afterwards)."""
-    rows = _websites_raw(ws)
-    if rows is None:
-        return []
-    out = [{"host": r.get("host"), "mode": _canon_mode(r.get("mode"))}
-           for r in rows]
-    _back_to_root(ws)
-    return out
-
-
-def website_mode(ws, host):
-    """Mode for one host, or None when absent/unreadable."""
-    for row in websites(ws):
-        if (row.get("host") or "").lower() == host.lower():
-            return row.get("mode")
-    return None
-
-
-def set_website_mode(ws, host, mode):
-    """Set one host mode via the mode chooser. Bool.
-
-    One-way for Ask/Deny: the override row leaves the allowed list
-    (no add UI), so removal verifies by absence. No-op when already
-    there; absent hosts fail (nothing to click).
-    """
-    if mode not in WEBSITE_MODES:
-        return False
-    rows = _websites_raw(ws)
-    if rows is None:
-        return False
-    target = next((r for r in rows
-                   if (r.get("host") or "").lower() == host.lower()),
-                  None)
-    if target is None:
-        _back_to_root(ws)
-        return False
-    if _canon_mode(target.get("mode")) == mode:
-        _back_to_root(ws)
-        return True
-    try:
-        real_click(ws, target["x"], target["y"])
-    except Exception:
-        _back_to_root(ws)
-        return False
-    time.sleep(0.8)
-    items = _eval(ws, JS_MODE_MENU)
-    texts = [(i.get("text") or "") for i in items] \
-        if isinstance(items, list) else []
-    if mode not in texts:
-        escape(ws)
-        _back_to_root(ws)
-        return False
-    if _eval(ws, JS_CLICK_MODE % mode) != "CLICKED":
-        escape(ws)
-        _back_to_root(ws)
-        return False
-    for _ in range(5):
-        time.sleep(2.0)
-        rows = _eval(ws, JS_WEBSITES)
-        if not isinstance(rows, list):
-            continue
-        cur = next((_canon_mode(r.get("mode")) for r in rows
-                    if (r.get("host") or "").lower() == host.lower()),
-                   None)
-        if mode in ("Ask", "Deny"):
-            if cur is None:
-                _back_to_root(ws)
-                return True
-        elif cur == mode:
-            _back_to_root(ws)
-            return True
-    escape(ws)
-    _back_to_root(ws)
-    return False
-
-
-def _protocols_raw(ws):
-    """Drill into protocols; rows or None (stays on sub-page)."""
-    if not dialog.click_row(ws, "Direct network protocols", TAB):
-        return None
-    return _stable_rows(ws, JS_PROTO_ROWS)
-
-
-def protocols(ws):
-    """[{slug, title, on}] (back at root afterwards)."""
-    rows = _protocols_raw(ws)
-    if rows is None:
-        return []
-    out = [{"slug": _slug(r.get("title", "")),
-            "title": r.get("title", ""),
-            "on": "on" if r.get("checked") else "off"} for r in rows]
-    _back_to_root(ws)
-    return out
-
-
-def protocol_state(ws, title):
-    """on/off for one protocol row title, None when absent."""
-    for row in protocols(ws):
-        if row.get("title") == title:
-            return row.get("on")
-    return None
-
-
-def set_protocol(ws, title, on):
-    """Set one protocol switch in place; readback before returning."""
-    rows = _protocols_raw(ws)
-    if rows is None:
-        return False
-    target = next((r for r in rows if r.get("title") == title), None)
-    if target is None:
-        _back_to_root(ws)
-        return False
-    want = bool(on)
-    if bool(target.get("checked")) == want:
-        _back_to_root(ws)
-        return True
-    try:
-        real_click(ws, target["x"], target["y"])
-    except Exception:
-        _back_to_root(ws)
-        return False
-    for _ in range(8):
-        time.sleep(2.0)
-        rows = _eval(ws, JS_PROTO_ROWS)
-        if not isinstance(rows, list):
-            continue
-        cur = next((r for r in rows if r.get("title") == title), None)
-        if cur is not None and bool(cur.get("checked")) == want:
-            _back_to_root(ws)
-            return True
-    _back_to_root(ws)
-    # In-dialog verify missed (slow commit or commit-on-close); the
-    # toggles-level fresh readback is the source of truth.
-    return False
-
-
-def manage_counts(ws):
-    """Manageable-row summary counts (name -> count|None)."""
-    if not dialog.goto_tab(ws, TAB):
-        return {}
-    text = dialog.dialog_text(ws) or ""
-    out = {}
-    for name in ("Websites", "Connectors", "Scheduled tasks",
-                 "Direct network protocols"):
-        m = re.search(re.escape(name) + r"\s*(\d+)", text)
-        out[name] = int(m.group(1)) if m else None
-    return out
-
-
-def scheduled_tasks(ws):
-    """[{name, cadence}] (back at root afterwards; empty when none)."""
-    if not dialog.click_row(ws, "Scheduled tasks", TAB):
-        return []
-    time.sleep(0.6)
-    text = dialog.dialog_text(ws) or ""
-    _back_to_root(ws)
-    rows = []
-    lines = [line.strip() for line in text.splitlines()
-             if line.strip()]
-    for i, line in enumerate(lines):
-        if re.search(r"\b(daily|weekly|hourly|every|min)\b", line,
-                     re.IGNORECASE) and i > 0:
-            rows.append({"name": lines[i - 1], "cadence": line})
-    return rows
-
-
-def all_toggles(ws):
-    """Flat map of every settable Permissions toggle (for list)."""
-    out = {}
-    defs = defaults(ws)
-    out["permissions.connector_defaults"] = defs.get("connector_defaults")
-    out["permissions.web_access"] = defs.get("web_access")
-    adv = advanced(ws)
-    for key, val in adv.items():
-        out["permissions.advanced." + key] = val
-    for row in protocols(ws):
-        if row.get("slug"):
-            out["permissions.protocols:" + row["slug"]] = row.get("on")
-    for row in websites(ws):
-        if row.get("host"):
-            out["permissions.websites:" + row["host"]] = row.get("mode")
-    return out
-
-
-def describe(ws):
-    """Full Permissions inventory: defaults, counts, adv, rows."""
-    return {"defaults": defaults(ws), "counts": manage_counts(ws),
-            "advanced": advanced(ws), "websites": websites(ws),
-            "protocols": protocols(ws),
-            "scheduled_tasks": scheduled_tasks(ws)}

bin/hatch_menu/tabs/secure_store.py

@@ -1,10 +0,0 @@
-"""Secure store tab: read-only inventory (secret entries + add)."""
-from hatch_menu import dialog
-
-TAB = "Secure store"
-
-
-def describe(ws):
-    """Row inventory + text excerpt (describe-only for now)."""
-    return {"rows": dialog.describe_rows(ws, TAB),
-            "text": (dialog.dialog_text(ws) or "")[:400]}

bin/hatch_menu/tabs/wallet.py

@@ -1,10 +0,0 @@
-"""Wallet tab: read-only inventory (payment methods + add row)."""
-from hatch_menu import dialog
-
-TAB = "Wallet"
-
-
-def describe(ws):
-    """Row inventory + text excerpt (describe-only for now)."""
-    return {"rows": dialog.describe_rows(ws, TAB),
-            "text": (dialog.dialog_text(ws) or "")[:400]}

bin/hatch_menu/toggles.py

@@ -1,319 +0,0 @@
-"""Toggle registry + sessions over controls.py and tabs/*.
-
-Every settable toggle has ONE address; contracts live in the tab
-modules (single-copy per part), addressing here. Set flows read back
-through a fresh session and never partially report success.
-
-Caller errors (unknown node/toggle/value) raise MenuError before any
-CDP traffic. Transport failures return {"ok": False, ...}.
-"""
-from approvals import VALID_NODES, get_cdp_ws
-from hatch_menu import controls, dialog
-from hatch_menu.mouse import close, escape
-from hatch_menu.tabs import TAB_MODULES
-
-_PERM = TAB_MODULES["Permissions"]
-_DC = TAB_MODULES["Data controls"]
-_GEN = TAB_MODULES["General"]
-
-
-class MenuError(ValueError):
-    """Caller error: unknown node, toggle, or value. Raised before CDP."""
-
-
-TOGGLES = {
-    "permissions.connector_defaults": {
-        "tab": "Permissions", "kind": "radio-heading",
-        "heading": _PERM.CONNECTOR_HEADING,
-        "values": _PERM.DEFAULT_VALUES},
-    "permissions.web_access": {
-        "tab": "Permissions", "kind": "radio-heading",
-        "heading": _PERM.WEB_HEADING,
-        "values": _PERM.DEFAULT_VALUES},
-    "permissions.advanced.transparent_proxy": {
-        "tab": "Permissions", "kind": "adv-switch",
-        "label": _PERM.ADV_LABELS["transparent_proxy"],
-        "values": ("on", "off")},
-    "permissions.advanced.tls_interception": {
-        "tab": "Permissions", "kind": "adv-switch",
-        "label": _PERM.ADV_LABELS["tls_interception"],
-        "values": ("on", "off")},
-    "permissions.advanced.sni_mismatch_rejection": {
-        "tab": "Permissions", "kind": "adv-switch",
-        "label": _PERM.ADV_LABELS["sni_mismatch_rejection"],
-        "values": ("on", "off")},
-    "data_controls.ai_improvement": {
-        "tab": "Data controls", "kind": "switch",
-        "values": ("on", "off"),
-        # Live 2026-10-06: the site ignores every input gesture here
-        # (synthetic/real/double/hover/keyboard/drag) — reads fine.
-        "readonly": True},
-    "general.theme": {
-        "tab": "General", "kind": "radio-aria",
-        "values": _GEN.THEME_VALUES},
-}
-
-WEBSITE_PREFIX = "permissions.websites:"
-PROTOCOL_PREFIX = "permissions.protocols:"
-
-
-def _check_node(node):
-    if node not in VALID_NODES:
-        raise MenuError("unknown node: %r (valid: %s)"
-                        % (node, ", ".join(VALID_NODES)))
-
-
-def normalize_onoff(value):
-    """on/off/true/false/1/0/yes/no -> bool. None when invalid."""
-    if isinstance(value, bool):
-        return value
-    if not isinstance(value, str):
-        return None
-    v = value.strip().lower()
-    if v in ("on", "true", "1", "yes"):
-        return True
-    if v in ("off", "false", "0", "no"):
-        return False
-    return None
-
-
-def resolve_toggle(name):
-    """Resolve a toggle address to its spec. Raises MenuError."""
-    if not isinstance(name, str) or not name:
-        raise MenuError("toggle name must be a non-empty string")
-    if name in TOGGLES:
-        spec = dict(TOGGLES[name])
-        spec["name"] = name
-        return spec
-    if name.startswith(WEBSITE_PREFIX) and len(name) > len(WEBSITE_PREFIX):
-        return {"name": name, "tab": "Permissions", "kind": "website",
-                "host": name[len(WEBSITE_PREFIX):],
-                "values": _PERM.WEBSITE_MODES}
-    if name.startswith(PROTOCOL_PREFIX) and len(name) > len(PROTOCOL_PREFIX):
-        label = _PERM.resolve_protocol(name[len(PROTOCOL_PREFIX):])
-        if label is None:
-            raise MenuError("unknown protocol: %r (see list_toggles)"
-                            % (name,))
-        return {"name": name, "tab": "Permissions", "kind": "protocol",
-                "label": label, "values": ("on", "off")}
-    raise MenuError("unknown toggle: %r (see list_toggles)" % (name,))
-
-
-def _session(node):
-    _check_node(node)
-    try:
-        ws, _ = get_cdp_ws(node)
-    except Exception as e:
-        raise MenuError("CDP unreachable for %s: %s" % (node, e))
-    return ws
-
-
-def _normalize_value(spec, value):
-    kind = spec["kind"]
-    if kind == "website":
-        if not isinstance(value, str):
-            raise MenuError("mode must be one of %s"
-                            % (spec["values"],))
-        for m in spec["values"]:
-            if value.strip().lower() == m.lower():
-                return m
-        raise MenuError("mode must be one of %s (got %r)"
-                        % (spec["values"], value))
-    if kind in ("switch", "adv-switch", "protocol"):
-        b = normalize_onoff(value)
-        if b is None:
-            raise MenuError("value must be on/off (got %r)" % (value,))
-        return "on" if b else "off"
-    if kind == "radio-heading":
-        if value not in spec["values"]:
-            raise MenuError("%s must be one of %s (got %r)"
-                            % (spec["name"], spec["values"], value))
-        return value
-    if kind == "radio-aria":
-        if not isinstance(value, str):
-            raise MenuError("theme must be one of %s" % (spec["values"],))
-        v = value.strip().lower()
-        if v in spec["values"]:
-            return v
-        raise MenuError("theme must be one of %s (got %r)"
-                        % (spec["values"], value))
-    raise MenuError("cannot set kind %r" % (kind,))
-
-
-def get_toggle(node, name):
-    """Read one toggle. Returns {"ok", "node", "toggle", "value"}."""
-    spec = resolve_toggle(name)
-    _check_node(node)
-    try:
-        ws = _session(node)
-    except MenuError as e:
-        return {"ok": False, "node": node, "toggle": name,
-                "error": str(e)}
-    try:
-        if not dialog.open_settings(ws):
-            return {"ok": False, "node": node, "toggle": name,
-                    "error": "settings dialog did not open"}
-        if not dialog.goto_tab(ws, spec["tab"]):
-            return {"ok": False, "node": node, "toggle": name,
-                    "error": "tab did not open: %s" % spec["tab"]}
-        kind = spec["kind"]
-        if kind == "radio-heading":
-            vals = {r["value"]: r["checked"]
-                    for r in controls.list_radios(ws)
-                    if (r.get("heading") or "").lower()
-                    == spec["heading"].lower()}
-            on = [v for v, c in vals.items() if c]
-            value = on[0] if len(on) == 1 else None
-        elif kind == "radio-aria":
-            vals = [(r.get("value"), r.get("checked"))
-                    for r in controls.list_radios(ws)]
-            on = [v for v, c in vals if c]
-            value = on[0] if len(on) == 1 else None
-        elif kind == "switch":
-            state = _DC.ai_improvement(ws)
-            value = ("on" if state else "off") if state is not None \
-                else None
-        elif kind == "adv-switch":
-            _PERM.ensure_advanced(ws)
-            state = controls.switch_state(ws, spec["label"])
-            value = ("on" if state else "off") if state is not None \
-                else None
-        elif kind == "website":
-            value = _PERM.website_mode(ws, spec["host"])
-        elif kind == "protocol":
-            value = _PERM.protocol_state(ws, spec["label"])
-        else:
-            value = None
-        if value is None:
-            if kind == "website":
-                return {"ok": False, "node": node, "toggle": name,
-                        "error": "host not in Websites list (effective: "
-                        "permissions.web_access default)"}
-            return {"ok": False, "node": node, "toggle": name,
-                    "error": "toggle not readable (site changed?)"}
-        return {"ok": True, "node": node, "toggle": name, "value": value}
-    except Exception as e:
-        return {"ok": False, "node": node, "toggle": name,
-                "error": "%s: %s" % (type(e).__name__, e)}
-    finally:
-        escape(ws)
-        close(ws)
-
-
-def set_toggle(node, name, value):
-    """Set one toggle with readback. Never partially reports success."""
-    spec = resolve_toggle(name)
-    if spec.get("readonly"):
-        raise MenuError("%s is read-only: the site ignores all input "
-                        "gestures there (locked?)" % name)
-    want = _normalize_value(spec, value)
-    _check_node(node)
-    try:
-        ws = _session(node)
-    except MenuError as e:
-        return {"ok": False, "node": node, "toggle": name,
-                "error": str(e)}
-    try:
-        if not dialog.open_settings(ws):
-            return {"ok": False, "node": node, "toggle": name,
-                    "error": "settings dialog did not open"}
-        if not dialog.goto_tab(ws, spec["tab"]):
-            return {"ok": False, "node": node, "toggle": name,
-                    "error": "tab did not open: %s" % spec["tab"]}
-        kind = spec["kind"]
-        if kind == "radio-heading":
-            ok = controls.set_radio_by_heading(ws, spec["heading"], want)
-        elif kind == "radio-aria":
-            ok = controls.set_radio_by_aria(ws, want)
-        elif kind == "switch":
-            ok = _DC.set_ai_improvement(ws, want == "on")
-        elif kind == "adv-switch":
-            _PERM.ensure_advanced(ws)
-            ok = controls.set_switch(ws, spec["label"], want == "on")
-        elif kind == "website":
-            ok = _PERM.set_website_mode(ws, spec["host"], want)
-        elif kind == "protocol":
-            ok = _PERM.set_protocol(ws, spec["label"], want == "on")
-        else:
-            ok = False
-        # The fresh-session readback is the source of truth: switch
-        # commits can land slowly or on dialog close, after the
-        # in-flow verify had its chance.
-        readback = get_toggle(node, name)
-        if readback.get("ok") and readback.get("value") == want:
-            out = {"ok": True, "node": node, "toggle": name,
-                   "value": want}
-            if not ok:
-                out["readback_only"] = True
-            return out
-        return {"ok": False, "node": node, "toggle": name,
-                "error": "readback mismatch (want %r, got %r)"
-                % (want, readback.get("value"))}
-    except Exception as e:
-        return {"ok": False, "node": node, "toggle": name,
-                "error": "%s: %s" % (type(e).__name__, e)}
-    finally:
-        escape(ws)
-        close(ws)
-
-
-def list_toggles(node, tab=None):
-    """All toggle states, optionally filtered to one tab."""
-    _check_node(node)
-    if tab is not None and tab not in TAB_MODULES:
-        raise MenuError("unknown tab: %r (valid: %s)"
-                        % (tab, sorted(TAB_MODULES)))
-    want = [tab] if tab else ["Permissions", "Data controls", "General"]
-    try:
-        ws = _session(node)
-    except MenuError as e:
-        return {"ok": False, "node": node, "error": str(e)}
-    try:
-        if not dialog.open_settings(ws):
-            return {"ok": False, "node": node,
-                    "error": "settings dialog did not open"}
-        out = {}
-        if "Permissions" in want:
-            dialog.goto_tab(ws, "Permissions")
-            out.update(_PERM.all_toggles(ws))
-        if "Data controls" in want:
-            state = _DC.ai_improvement(ws)
-            out["data_controls.ai_improvement"] = \
-                ("on" if state else "off") if state is not None else None
-        if "General" in want:
-            dialog.goto_tab(ws, "General")
-            out["general.theme"] = _GEN.theme(ws)
-        return {"ok": True, "node": node, "toggles": out}
-    except Exception as e:
-        return {"ok": False, "node": node,
-                "error": "%s: %s" % (type(e).__name__, e)}
-    finally:
-        escape(ws)
-        close(ws)
-
-
-def describe_tab(node, tab):
-    """Full inventory of one tab (debugging/patching aid)."""
-    _check_node(node)
-    if tab not in TAB_MODULES:
-        raise MenuError("unknown tab: %r (valid: %s)"
-                        % (tab, sorted(TAB_MODULES)))
-    try:
-        ws = _session(node)
-    except MenuError as e:
-        return {"ok": False, "node": node, "tab": tab, "error": str(e)}
-    try:
-        if not dialog.open_settings(ws):
-            return {"ok": False, "node": node, "tab": tab,
-                    "error": "settings dialog did not open"}
-        if not dialog.goto_tab(ws, tab):
-            return {"ok": False, "node": node, "tab": tab,
-                    "error": "tab did not open"}
-        return {"ok": True, "node": node, "tab": tab,
-                "inventory": TAB_MODULES[tab].describe(ws)}
-    except Exception as e:
-        return {"ok": False, "node": node, "tab": tab,
-                "error": "%s: %s" % (type(e).__name__, e)}
-    finally:
-        escape(ws)
-        close(ws)

bin/host_evidence.py

@@ -1,348 +0,0 @@
-#!/usr/bin/env python3
-"""Host-side fleet evidence for network/PID-blind shells.
-
-`box fleet status` probes each node live (pgrep for the chromium process,
-HTTP to the CDP relay on the peer IP). Both probes assume the caller's
-network + PID namespace is the bl host's. From a sandboxed shell (own PID
-and net namespaces, no sudo, no route to 10.201.x.x) both probes always
-fail, so every node misreports as STOPPED even with a healthy fleet.
-
-This module provides the fallback signal: evidence written by the
-host-side watchdogs that run on bl unsandboxed via systemd timers:
-
-  - cdp-relay-watchdog (every 5 min, all active registry nodes):
-    log ``cdp-relay-watchdog.log`` + journal unit
-    ``cdp-relay-watchdog.service``. Proves the CDP relay path end to end.
-  - chromebox-watchdog (every 2 min, same nodes, one timer per node):
-    log ``chromebox-watchdog.log`` + journal units
-    ``chromebox-watchdog@<node>.service``. Curls CDP inside the node netns,
-    so it proves browser + in-netns CDP.
-  - ``chromebox-<node>.log`` mtime: chromium's own stdout. Fresh output
-    proves the browser process is alive. Used only for nodes outside
-    watchdog coverage — and only to conclude "alive", never "dead".
-
-Both watchdogs are silent-when-healthy: a failure is ALWAYS logged, so a
-recent timer run (journal "Starting" line) with no newer failure line for
-the node means that run found the node healthy.
-
-Verdicts: "healthy" | "degraded" | "down" | "unknown".
-"""
-
-import json
-import os
-import re
-import subprocess
-import time
-from datetime import datetime, timezone
-from pathlib import Path
-
-NETVM_ROOT = Path("/home/super/Projects/NetVM")
-RELAY_LOG = NETVM_ROOT / "cdp-relay-watchdog.log"
-CHROMEBOX_LOG = NETVM_ROOT / "chromebox-watchdog.log"
-
-ALL_NODES = ("muse", "pip", "646", "opm", "def", "dev")
-
-
-def _covered_nodes():
-    """Nodes with watchdog coverage, from the fleet registry.
-
-    Both watchdogs supervise every active registry node. Falls back to
-    ALL_NODES when the registry is unreadable, so a broken registry can
-    never silently narrow fleet status to a subset of the fleet.
-    """
-    try:
-        import importlib.util
-        spec = importlib.util.spec_from_file_location(
-            "netvm_registry", NETVM_ROOT / "bin" / "netvm-registry.py")
-        mod = importlib.util.module_from_spec(spec)
-        spec.loader.exec_module(mod)
-        return tuple(sorted(mod.active_nodes()))
-    except Exception:
-        return ALL_NODES
-
-
-RELAY_NODES = _covered_nodes()
-CHROMEBOX_NODES = _covered_nodes()
-
-RELAY_UNIT = "cdp-relay-watchdog.service"
-CHROMEBOX_UNIT_TMPL = "chromebox-watchdog@{node}.service"
-
-# A watchdog run older than this proves nothing (timer may be dead).
-RELAY_STALE_MIN = 15
-CHROMEBOX_STALE_MIN = 8
-# Chromium stdout older than this proves nothing (idle browsers go quiet).
-CHROME_LOG_FRESH_MIN = 20
-
-_LOG_TS_RE = re.compile(r"^\[(\d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2})Z\]")
-
-
-def _utcnow():
-    return datetime.now(timezone.utc)
-
-
-def parse_log_ts(line):
-    """Parse a ``[YYYY-MM-DDTHH:MM:SSZ]`` log prefix. None if absent."""
-    m = _LOG_TS_RE.match(line)
-    if not m:
-        return None
-    try:
-        return datetime.strptime(m.group(1), "%Y-%m-%dT%H:%M:%S").replace(
-            tzinfo=timezone.utc)
-    except ValueError:
-        return None
-
-
-def classify_chromebox_line(line):
-    """Classify one chromebox-watchdog log line.
-
-    Returns "healthy" | "degraded" | "down", or None when the line
-    carries no verdict (rotation markers, relay stdout passthrough).
-    """
-    if "log rotated" in line:
-        return None
-    if "relaunch FAILED" in line:
-        return "down"
-    if "relaunch OK" in line:
-        return "healthy"
-    if "recovered" in line and "relaunch not needed" in line:
-        return "healthy"
-    if "not healthy yet" in line:
-        return "degraded"
-    if "proceeding with chrome relaunch" in line:
-        return "degraded"
-    if "relaunching chromebox" in line:
-        return "degraded"
-    if "warp partition detected" in line:
-        return "degraded"
-    if "skipping relaunch (probably still starting)" in line:
-        return "degraded"
-    return None
-
-
-def classify_relay_line(line):
-    """Classify one cdp-relay-watchdog log line (None = no verdict)."""
-    if "relay restart FAILED" in line:
-        return "down"
-    if "FAIL_LOUD" in line:
-        return "down"
-    if "relay restarted OK" in line:
-        return "healthy"
-    if "relay unhealthy" in line and "restarting" in line:
-        # Always followed by an OK/FAILED line; a trailing one means the
-        # restart crashed mid-flight.
-        return "down"
-    return None
-
-
-def last_verdict(lines, node, classify):
-    """Newest (verdict, ts, line) for ``[node]``. None if no verdict line."""
-    tag = "[%s]" % node
-    best = None
-    for line in lines:
-        if tag not in line:
-            continue
-        verdict = classify(line)
-        if verdict is None:
-            continue
-        ts = parse_log_ts(line)
-        if ts is None:
-            continue
-        if best is None or ts >= best[1]:
-            best = (verdict, ts, line.strip()[:160])
-    return best
-
-
-def _tail_lines(path, max_bytes=65536):
-    try:
-        size = os.path.getsize(path)
-        with open(path, "rb") as f:
-            if size > max_bytes:
-                f.seek(size - max_bytes)
-                f.readline()  # drop partial first line
-            return f.read().decode("utf-8", errors="replace").splitlines()
-    except OSError:
-        return []
-
-
-def query_journal_starts(units, since_min=25, timeout=20):
-    """Map each unit -> newest run-start (aware UTC). Missing on failure.
-
-    Uses ``-o json``: the short-format "Starting" line carries the unit
-    description, not the unit name, so exact per-unit matching needs the
-    structured UNIT field.
-    """
-    cmd = ["journalctl", "--no-pager", "-o", "json",
-           "--since", "%d min ago" % since_min]
-    for u in units:
-        cmd.extend(["-u", u])
-    try:
-        r = subprocess.run(cmd, capture_output=True, text=True, timeout=timeout)
-    except (OSError, subprocess.TimeoutExpired):
-        return {}
-    if r.returncode != 0:
-        return {}
-    want = set(units)
-    starts = {}
-    for line in (r.stdout or "").splitlines():
-        try:
-            e = json.loads(line)
-        except ValueError:
-            continue
-        if e.get("UNIT") not in want:
-            continue
-        if not (e.get("MESSAGE") or "").startswith("Starting"):
-            continue
-        try:
-            ts = datetime.fromtimestamp(
-                int(e["__REALTIME_TIMESTAMP"]) / 1e6, tz=timezone.utc)
-        except (KeyError, ValueError, TypeError, OverflowError):
-            continue
-        u = e["UNIT"]
-        if u not in starts or ts > starts[u]:
-            starts[u] = ts
-    return starts
-
-
-def _verdict_since_run(verdict_row, run_ts):
-    """True when the verdict line is newer than (or from) the last run."""
-    if verdict_row is None or run_ts is None:
-        return False
-    return verdict_row[1] >= run_ts
-
-
-def browser_verdict(node, chromebox_lines, run_ts, chrome_log_mtime=None,
-                    now=None):
-    """(verdict, detail) for the node's browser process."""
-    now = now or _utcnow()
-    row = last_verdict(chromebox_lines, node, classify_chromebox_line)
-    if node in CHROMEBOX_NODES:
-        if run_ts is None:
-            return ("unknown", "no chromebox-watchdog run in journal window")
-        if (now - run_ts).total_seconds() > CHROMEBOX_STALE_MIN * 60:
-            return ("unknown", "chromebox-watchdog run is stale")
-        if _verdict_since_run(row, run_ts):
-            return (row[0], "watchdog: %s" % row[2])
-        return ("healthy", "watchdog run silent (silent-when-healthy)")
-    # Nodes outside watchdog coverage: chromium stdout proves alive only.
-    if chrome_log_mtime is not None and (
-            now - chrome_log_mtime).total_seconds() < CHROME_LOG_FRESH_MIN * 60:
-        return ("healthy", "chromebox-%s.log fresh" % node)
-    return ("unknown", "no watchdog coverage for %s" % node)
-
-
-def cdp_verdict(node, relay_lines, run_ts, now=None):
-    """(verdict, detail) for the node's host-reachable CDP relay path."""
-    now = now or _utcnow()
-    if node not in RELAY_NODES:
-        return ("unknown", "no relay-monitor coverage for %s" % node)
-    if run_ts is None:
-        return ("unknown", "no cdp-relay-watchdog run in journal window")
-    if (now - run_ts).total_seconds() > RELAY_STALE_MIN * 60:
-        return ("unknown", "cdp-relay-watchdog run is stale")
-    row = last_verdict(relay_lines, node, classify_relay_line)
-    if _verdict_since_run(row, run_ts):
-        return (row[0], "relay watchdog: %s" % row[2])
-    return ("healthy", "relay watchdog run silent (silent-when-healthy)")
-
-
-def chrome_log_mtime(node):
-    """Mtime of chromium's stdout log as aware UTC. None if missing."""
-    try:
-        return datetime.fromtimestamp(
-            os.path.getmtime(NETVM_ROOT / ("chromebox-%s.log" % node)),
-            tz=timezone.utc)
-    except OSError:
-        return None
-
-
-_CACHE = {"at": 0.0, "nodes": frozenset(), "data": {}}
-_CACHE_TTL_S = 60
-
-
-def collect(nodes=None, _journal_starts=None, _relay_lines=None,
-            _chromebox_lines=None, _chrome_mtimes=None, _now=None):
-    """Per-node host evidence. Underscore args are seams for tests."""
-    nodes = list(nodes or ALL_NODES)
-    live = (_journal_starts is None and _relay_lines is None
-            and _chromebox_lines is None and _chrome_mtimes is None
-            and _now is None)
-    if live:
-        key = frozenset(nodes)
-        if (key <= _CACHE["nodes"]
-                and time.monotonic() - _CACHE["at"] < _CACHE_TTL_S):
-            return {n: _CACHE["data"][n] for n in nodes if n in _CACHE["data"]}
-    now = _now or _utcnow()
-    if _journal_starts is None:
-        units = [RELAY_UNIT] + [CHROMEBOX_UNIT_TMPL.format(node=n)
-                                for n in nodes if n in CHROMEBOX_NODES]
-        _journal_starts = query_journal_starts(units)
-    if _relay_lines is None:
-        _relay_lines = _tail_lines(RELAY_LOG)
-    if _chromebox_lines is None:
-        _chromebox_lines = _tail_lines(CHROMEBOX_LOG)
-    out = {}
-    for node in nodes:
-        if _chrome_mtimes is not None and node in _chrome_mtimes:
-            mtime = _chrome_mtimes[node]
-        else:
-            mtime = chrome_log_mtime(node) if node not in CHROMEBOX_NODES else None
-        b_verd, b_det = browser_verdict(
-            node, _chromebox_lines,
-            _journal_starts.get(CHROMEBOX_UNIT_TMPL.format(node=node)),
-            chrome_log_mtime=mtime, now=now)
-        c_verd, c_det = cdp_verdict(
-            node, _relay_lines, _journal_starts.get(RELAY_UNIT), now=now)
-        out[node] = {
-            "browser": b_verd,
-            "browser_detail": b_det,
-            "cdp": c_verd,
-            "cdp_detail": c_det,
-        }
-    if live:
-        _CACHE["at"] = time.monotonic()
-        _CACHE["nodes"] = frozenset(nodes)
-        _CACHE["data"] = out
-    return out
-
-
-def effective_status(local_proc, local_cdp, browser_v, cdp_v):
-    """Map (local probes, host verdicts) -> (status, source).
-
-    Host evidence only ever overrides the fully-blind pattern (both
-    local probes negative — the sandbox signature). It never overrides
-    a live local signal, so a fresh outage on the host always wins.
-    """
-    if local_cdp:
-        # CDP answers: the browser is definitionally alive.
-        return ("ACTIVE", "local")
-    if local_proc:
-        return ("CDP_DOWN", "local")
-    # Both local probes negative: consult host evidence.
-    if browser_v == "down":
-        return ("STOPPED", "host-evidence")
-    if browser_v == "unknown":
-        return ("UNKNOWN", "host-evidence")
-    if cdp_v == "healthy":
-        return ("ACTIVE", "host-evidence")
-    if cdp_v == "down":
-        return ("CDP_DOWN", "host-evidence")
-    return ("UNKNOWN", "host-evidence")
-
-
-def main(argv=None):
-    import argparse
-    ap = argparse.ArgumentParser(description="Show host-side fleet evidence")
-    ap.add_argument("--json", action="store_true")
-    args = ap.parse_args(argv)
-    data = collect()
-    if args.json:
-        print(json.dumps({"ok": True, "evidence": data}, indent=2))
-        return
-    for node, ev in data.items():
-        print("%-6s browser=%-8s cdp=%-8s" % (node, ev["browser"], ev["cdp"]))
-        print("       browser: %s" % ev["browser_detail"])
-        print("       cdp:     %s" % ev["cdp_detail"])
-
-
-if __name__ == "__main__":
-    main()

bin/identity-audit-check.sh

@@ -1,70 +0,0 @@
-#!/bin/bash
-# identity-audit-check.sh - Check VM identity audit for drift
-#
-# Lives on bl (/home/super/Projects/NetVM/bin/). Reads a bl-local cached
-# copy of the VM's identity audit JSON and reports drift.
-#
-# Why a cache: bl cannot SSH to the VM (VM only accepts the operator
-# container's key). The VM's hourly audit cron (/srv/board/bin/identity-audit.py)
-# should scp /srv/board/data/identity-audit.json to bl at:
-#   /home/super/Projects/NetVM/var/identity-audit.json
-# after each run. Until that push is wired, the cache is refreshed manually
-# or by the identity-audit-watch cron.
-#
-# Called by:
-#   - the identity-audit-watch cron (replaces inline SSH one-liner)
-#   - box-ctl.py `identity-audit` action
-#
-# Exit codes:
-#   0 - clean (no drift; warnings are expected and silent)
-#   1 - drift detected (items printed to stdout, one per line)
-#   2 - audit cache missing/unreadable (needs attention)
-#
-# Output on drift: one drift item per line, prefixed with "DRIFT: "
-
-set -u
-
-CACHE_PATH="/home/super/Projects/NetVM/var/identity-audit.json"
-
-# Allow override for testing
-if [ $# -ge 1 ] && [ -f "$1" ]; then
-    CACHE_PATH="$1"
-fi
-
-if [ ! -f "$CACHE_PATH" ]; then
-    echo "ERROR: identity audit cache missing: $CACHE_PATH" >&2
-    echo "The VM hourly audit should push /srv/board/data/identity-audit.json here after each run." >&2
-    exit 2
-fi
-
-audit_json=$(cat "$CACHE_PATH" 2>/dev/null)
-if [ -z "$audit_json" ]; then
-    echo "ERROR: identity audit cache unreadable: $CACHE_PATH" >&2
-    exit 2
-fi
-
-# Parse with python3
-drift_output=$(printf '%s' "$audit_json" | python3 -c "
-import json, sys
-try:
-    data = json.load(sys.stdin)
-except Exception as e:
-    print('ERROR: invalid JSON in audit cache: %s' % e, file=sys.stderr)
-    sys.exit(2)
-for item in data.get('drift', []):
-    print('DRIFT: ' + str(item))
-" 2>&1)
-parse_rc=$?
-
-if [ $parse_rc -eq 2 ]; then
-    echo "$drift_output" >&2
-    exit 2
-fi
-
-if [ -n "$drift_output" ]; then
-    echo "$drift_output"
-    exit 1
-fi
-
-# Clean: no drift. Warnings are expected (agents not dialed in) — stay silent.
-exit 0

bin/identity-broker.py

@@ -1,342 +0,0 @@
-#!/usr/bin/env python3
-"""identity-broker.py — Scope lifecycle over proxy providers.
-
-Owns identity-state.json (gitignored runtime state: scope -> label
-assignments, no secrets) and drives providers through it:
-
-  up <fp>                  resolve fingerprint, provision its scope
-  down <fp|unit>           teardown scope network, drop assignment
-  cycle <fp>               rotate to a fresh identity (bumps cycles)
-  exec <fp> -- <cmd>       run a command inside the scope network
-  routes <fp>              read-only route/tunnel status
-  status                   scopes with emails/labels (no key material)
-  bind --from <runs.json> --session <s> --fp <f>
-                           attribute live runs (agent-manager --once
-                           --json) to a scope after verifying the
-                           session exists in the scan
-
-v1 boundary: the broker scopes NETWORK identity only. It never sees,
-stores, or prints key bytes — fingerprints and emails are the only
-identifiers here. Short-lived brokered credential issuance is a
-deferred stage; harnesses receive keys through existing means.
-"""
-
-from __future__ import annotations
-
-import argparse
-import importlib.util
-import json
-import sys
-import time
-from pathlib import Path
-from typing import Any, Callable, Dict, List, Optional, Tuple
-
-REPO_ROOT = Path(__file__).resolve().parent.parent
-BIN_DIR = REPO_ROOT / "bin"
-STATE_FILE = REPO_ROOT / "identity-state.json"
-
-RunFn = Callable[..., Tuple[int, str]]
-
-
-def _load(name: str, modname: str):
-    path = BIN_DIR / name
-    spec = importlib.util.spec_from_file_location(modname, path)
-    mod = importlib.util.module_from_spec(spec)
-    sys.modules[modname] = mod
-    spec.loader.exec_module(mod)
-    return mod
-
-
-resolve_mod = _load("identity-resolve.py", "identity_resolve")
-provider_mod = _load("identity-provider.py", "identity_provider")
-
-PROVIDERS = {"warp": provider_mod.WarpProvider(),
-               "wireguard": provider_mod.GenericWireGuardProvider(),
-               "socks": provider_mod.SocksProxyProvider()}
-
-
-class BrokerError(RuntimeError):
-    """A broker operation failed (message is safe to show)."""
-
-
-def load_state(path: str | Path = STATE_FILE) -> Dict[str, Any]:
-    """Load broker state. Missing/corrupt -> empty (never raises)."""
-    try:
-        with open(path, "r") as f:
-            data = json.load(f)
-        if isinstance(data, dict):
-            data.setdefault("scopes", {})
-            data.setdefault("bindings", [])
-            return data
-    except Exception:
-        pass
-    return {"scopes": {}, "bindings": []}
-
-
-def save_state(state: Dict[str, Any],
-               path: str | Path = STATE_FILE) -> None:
-    with open(path, "w") as f:
-        json.dump(state, f, indent=2, sort_keys=True)
-
-
-def _provider(name: str = "warp"):
-    try:
-        prov = PROVIDERS[name]
-    except KeyError:
-        raise BrokerError("unknown provider %r (have: %s)"
-                          % (name, ", ".join(sorted(PROVIDERS))))
-    if not prov.ready:
-        raise BrokerError("provider %r is boilerplate (not implemented); "
-                          "warp is the live provider" % (name,))
-    return prov
-
-
-def _scope_for_fp(fp: str, map_path=None) -> Dict[str, Any]:
-    scope = resolve_mod.resolve_scope(
-        resolve_mod.load_map(map_path or resolve_mod.MAP_FILE), fp)
-    if scope is None:
-        raise BrokerError("unknown fingerprint (not in identity-map.json)")
-    scope["slug"] = resolve_mod.scope_slug(scope)
-    return scope
-
-
-def _unit_key(fp_or_unit: str, state: Dict[str, Any],
-              map_path=None) -> str:
-    """Resolve CLI input (fp or scope unit) to a state scopes key."""
-    scopes = state.get("scopes", {})
-    if fp_or_unit in scopes:
-        return fp_or_unit
-    try:
-        scope = _scope_for_fp(fp_or_unit, map_path)
-    except BrokerError:
-        raise BrokerError("no scope for %r (unknown fingerprint, no "
-                          "such unit)" % (fp_or_unit,))
-    if scope["unit"] not in scopes:
-        raise BrokerError("scope %s is not up" % (scope["unit"],))
-    return scope["unit"]
-
-
-def op_up(fp: str, run: Optional[RunFn] = None, map_path=None,
-          state_path: str | Path = STATE_FILE,
-          provider_name: str = "warp") -> Dict[str, Any]:
-    """Provision a scope's network. Idempotent (re-up returns existing)."""
-    scope = _scope_for_fp(fp, map_path)
-    state = load_state(state_path)
-    if scope["unit"] in state["scopes"]:
-        return {"ok": "exists", **state["scopes"][scope["unit"]]}
-    label = scope["slug"]
-    try:
-        res = _provider(provider_name).provision(label, run=run)
-    except provider_mod.ProviderError as e:
-        raise BrokerError(str(e))
-    rec = {"scope": scope["scope"], "email": scope["email"],
-           "origins": scope["origins"], "label": label,
-           "provider": provider_name, "netns": res.get("netns", ""),
-           "created": int(time.time()), "cycles": 0}
-    state["scopes"][scope["unit"]] = rec
-    save_state(state, state_path)
-    return {"ok": "true", **rec}
-
-
-def op_down(fp_or_unit: str, run: Optional[RunFn] = None, map_path=None,
-            state_path: str | Path = STATE_FILE) -> Dict[str, Any]:
-    """Teardown a scope's network and drop its assignment + bindings."""
-    state = load_state(state_path)
-    unit = _unit_key(fp_or_unit, state, map_path)
-    rec = state["scopes"][unit]
-    try:
-        _provider(rec.get("provider", "warp")).teardown(rec["label"],
-                                                        run=run)
-    except provider_mod.ProviderError as e:
-        raise BrokerError(str(e))
-    del state["scopes"][unit]
-    state["bindings"] = [b for b in state.get("bindings", [])
-                         if b.get("unit") != unit]
-    save_state(state, state_path)
-    return {"ok": "true", "unit": unit, "label": rec["label"]}
-
-
-def op_cycle(fp: str, run: Optional[RunFn] = None, map_path=None,
-             state_path: str | Path = STATE_FILE) -> Dict[str, Any]:
-    """Rotate a scope to a fresh identity (bumps the cycle count)."""
-    scope = _scope_for_fp(fp, map_path)
-    state = load_state(state_path)
-    if scope["unit"] not in state["scopes"]:
-        raise BrokerError("scope %s is not up (up it first)"
-                          % (scope["unit"],))
-    rec = state["scopes"][scope["unit"]]
-    try:
-        _provider(rec.get("provider", "warp")).cycle(rec["label"],
-                                                     run=run)
-    except provider_mod.ProviderError as e:
-        raise BrokerError(str(e))
-    rec["cycles"] = int(rec.get("cycles", 0)) + 1
-    save_state(state, state_path)
-    return {"ok": "true", "unit": scope["unit"], "label": rec["label"],
-            "cycles": rec["cycles"]}
-
-
-def op_exec(fp: str, cmd: List[str], run: Optional[RunFn] = None,
-            map_path=None,
-            state_path: str | Path = STATE_FILE) -> Tuple[int, str]:
-    """Run cmd inside the scope's network. Returns (rc, output)."""
-    scope = _scope_for_fp(fp, map_path)
-    state = load_state(state_path)
-    if scope["unit"] not in state["scopes"]:
-        raise BrokerError("scope %s is not up (up it first)"
-                          % (scope["unit"],))
-    rec = state["scopes"][scope["unit"]]
-    try:
-        return _provider(rec.get("provider", "warp")).exec(
-            rec["label"], cmd, run=run)
-    except provider_mod.ProviderError as e:
-        raise BrokerError(str(e))
-
-
-def op_routes(fp: str, run: Optional[RunFn] = None, map_path=None,
-              state_path: str | Path = STATE_FILE) -> Dict[str, Any]:
-    """Read-only route/tunnel status for a scope."""
-    scope = _scope_for_fp(fp, map_path)
-    state = load_state(state_path)
-    if scope["unit"] not in state["scopes"]:
-        raise BrokerError("scope %s is not up (up it first)"
-                          % (scope["unit"],))
-    rec = state["scopes"][scope["unit"]]
-    try:
-        info = _provider(rec.get("provider", "warp")).routes(
-            rec["label"], run=run)
-    except provider_mod.ProviderError as e:
-        raise BrokerError(str(e))
-    return {"unit": scope["unit"], "email": scope["email"], **info}
-
-
-def op_status(run: Optional[RunFn] = None,
-              state_path: str | Path = STATE_FILE) -> Dict[str, Any]:
-    """Scopes with live provider status. Emails/labels only."""
-    state = load_state(state_path)
-    scopes = []
-    for unit, rec in sorted(state.get("scopes", {}).items()):
-        try:
-            live = _provider(rec.get("provider", "warp")).status(
-                rec["label"], run=run)
-        except provider_mod.ProviderError as e:
-            live = {"conf": "?", "netns": "?", "egress": "n/a",
-                    "error": str(e)}
-        scopes.append({"unit": unit, "email": rec.get("email", ""),
-                       "scope": rec.get("scope", ""),
-                       "label": rec.get("label", ""),
-                       "provider": rec.get("provider", ""),
-                       "cycles": rec.get("cycles", 0), **live})
-    return {"scopes": scopes, "bindings": state.get("bindings", [])}
-
-
-def op_bind(runs_path: str, session: str, fp: str, map_path=None,
-            state_path: str | Path = STATE_FILE) -> Dict[str, Any]:
-    """Attribute live runs to a scope, verifying against a scan.
-
-    runs_path is agent-manager.py --once --json output. Every run
-    whose session group contains `session` is bound to the fp's scope
-    (fp must resolve; the scope need not be up — binding is
-    attribution, not network). Sessions absent from the scan are
-    refused (never bind what we cannot observe).
-    """
-    scope = _scope_for_fp(fp, map_path)
-    try:
-        with open(runs_path, "r") as f:
-            scan = json.load(f)
-        runs = scan.get("runs", [])
-        if not isinstance(runs, list):
-            raise ValueError("no runs list")
-    except Exception as e:
-        raise BrokerError("cannot read runs scan %s: %s" % (runs_path, e))
-    matched = []
-    for r in runs:
-        if not isinstance(r, dict):
-            continue
-        group = str(r.get("session", "")).split(",")
-        if session in group:
-            matched.append(r)
-    if not matched:
-        raise BrokerError("session %r not observed in %s (refusing to "
-                          "bind unseen runs)" % (session, runs_path))
-    state = load_state(state_path)
-    now = int(time.time())
-    new = []
-    for r in matched:
-        rec = {"unit": scope["unit"], "email": scope["email"],
-               "device": r.get("device", ""), "type": r.get("type", ""),
-               "session": session, "pane": r.get("pane", ""),
-               "pid": r.get("pid", 0), "bound_at": now}
-        new.append(rec)
-    # Replace prior bindings for these exact runs (re-bind refreshes).
-    keys = {(b["device"], b.get("pane"), b.get("pid")) for b in new}
-    state["bindings"] = [b for b in state.get("bindings", [])
-                         if (b.get("device"), b.get("pane"),
-                             b.get("pid")) not in keys] + new
-    save_state(state, state_path)
-    return {"ok": "true", "unit": scope["unit"], "bound": len(new),
-            "runs": new}
-
-
-def main(argv: Optional[List[str]] = None) -> int:
-    ap = argparse.ArgumentParser(prog="identity-broker.py")
-    ap.add_argument("--map", default=str(resolve_mod.MAP_FILE),
-                    help="identity map (default: identity-map.json)")
-    ap.add_argument("--state", default=str(STATE_FILE),
-                    help="broker state file (default: identity-state.json)")
-    sub = ap.add_subparsers(dest="cmd", required=True)
-    p = sub.add_parser("up", help="provision a scope network")
-    p.add_argument("fp")
-    p.add_argument("--provider", default="warp")
-    p = sub.add_parser("down", help="teardown a scope network")
-    p.add_argument("fp_or_unit")
-    p = sub.add_parser("cycle", help="rotate a scope identity")
-    p.add_argument("fp")
-    p = sub.add_parser("exec", help="run a command in a scope network")
-    p.add_argument("fp")
-    p.add_argument("exec_cmd", nargs=argparse.REMAINDER,
-                   help="command (after --)")
-    p = sub.add_parser("routes", help="route/tunnel status for a scope")
-    p.add_argument("fp")
-    sub.add_parser("status", help="scopes + bindings (emails only)")
-    p = sub.add_parser("bind", help="attribute live runs to a scope")
-    p.add_argument("--from", dest="runs", required=True)
-    p.add_argument("--session", required=True)
-    p.add_argument("--fp", required=True)
-    args = ap.parse_args(argv)
-    mp, sp = args.map, args.state
-    try:
-        if args.cmd == "up":
-            print(json.dumps(op_up(args.fp, map_path=mp,
-                                   state_path=sp,
-                                   provider_name=args.provider),
-                             indent=2))
-        elif args.cmd == "down":
-            print(json.dumps(op_down(args.fp_or_unit, map_path=mp,
-                                     state_path=sp), indent=2))
-        elif args.cmd == "cycle":
-            print(json.dumps(op_cycle(args.fp, map_path=mp,
-                                      state_path=sp), indent=2))
-        elif args.cmd == "exec":
-            cmd = [c for c in (args.exec_cmd or []) if c != "--"]
-            rc, out = op_exec(args.fp, cmd, map_path=mp,
-                              state_path=sp)
-            sys.stdout.write(out + ("\n" if out else ""))
-            return rc
-        elif args.cmd == "routes":
-            print(json.dumps(op_routes(args.fp, map_path=mp,
-                                       state_path=sp), indent=2))
-        elif args.cmd == "status":
-            print(json.dumps(op_status(state_path=sp), indent=2))
-        elif args.cmd == "bind":
-            print(json.dumps(op_bind(args.runs, args.session, args.fp,
-                                     map_path=mp, state_path=sp),
-                             indent=2))
-        return 0
-    except BrokerError as e:
-        print("error: %s" % e)
-        return 1
-
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/identity-provider.py

@@ -1,237 +0,0 @@
-#!/usr/bin/env python3
-"""identity-provider.py — Proxy provider implementations.
-
-A provider owns one network-identity substrate behind a fixed
-interface: provision / teardown / cycle / exec / routes / status.
-All subprocesses go through an injectable run function (same seam as
-box-fleet-tui gather_*), so command shapes are unit-testable and no
-test touches netns, sudo, or /etc/netvm.
-
-Security boundaries (from the repo's own scripts):
-- Warp identities generate via netvm-new-identity.sh, which the user
-  explicitly authorized operators to run (see netvm-provision-node.sh
-  header). Generation installs a root-0600 conf and prints nothing.
-- This code NEVER reads /etc/netvm and never prints key material.
-  Confs are consumed only by root tools (wg setconf inside netns).
-- CLI-facing output carries emails, labels, and fingerprints only.
-"""
-
-from __future__ import annotations
-
-import os
-import re
-import subprocess
-import sys
-from pathlib import Path
-from typing import Callable, Dict, List, Optional, Tuple
-
-REPO_ROOT = Path(__file__).resolve().parent.parent
-BIN_DIR = REPO_ROOT / "bin"
-
-RunFn = Callable[..., Tuple[int, str]]
-
-LABEL_RE = re.compile(r"^[a-z0-9][a-z0-9-]{0,22}$")
-
-
-def _run(cmd: List[str], timeout: int = 120) -> Tuple[int, str]:
-    """Run cmd, capture output. Returns (returncode, combined_output)."""
-    try:
-        r = subprocess.run(cmd, capture_output=True, text=True,
-                           timeout=timeout)
-        return r.returncode, ((r.stdout or "") + (r.stderr or "")).strip()
-    except subprocess.TimeoutExpired:
-        return 124, "timed out after %ds: %s" % (timeout, " ".join(cmd))
-    except OSError as e:
-        return 127, str(e)
-
-
-class ProviderError(RuntimeError):
-    """A provider operation failed (message is safe to show)."""
-
-
-def check_label(label: str) -> str:
-    """Validate a netvm label. Returns it or raises ProviderError."""
-    if not LABEL_RE.match(label or ""):
-        raise ProviderError(
-            "invalid label %r: lowercase letters, digits, hyphens "
-            "(max 23 chars)" % (label,))
-    return label
-
-
-class Provider:
-    """Interface every proxy provider implements. Boilerplate subclasses
-    override these with real substrate calls; see WarpProvider."""
-
-    name = "base"
-    ready = False
-
-    def provision(self, label: str,
-                  run: Optional[RunFn] = None) -> Dict[str, str]:
-        """Create the network identity + bring it up. Idempotent."""
-        raise NotImplementedError
-
-    def teardown(self, label: str,
-                 run: Optional[RunFn] = None) -> Dict[str, str]:
-        """Bring the identity's network down (keeps the identity)."""
-        raise NotImplementedError
-
-    def cycle(self, label: str,
-              run: Optional[RunFn] = None) -> Dict[str, str]:
-        """Rotate to a fresh identity (teardown + new identity + up)."""
-        raise NotImplementedError
-
-    def exec(self, label: str, cmd: List[str],
-             run: Optional[RunFn] = None) -> Tuple[int, str]:
-        """Run cmd inside the identity's network. Returns (rc, output)."""
-        raise NotImplementedError
-
-    def routes(self, label: str,
-               run: Optional[RunFn] = None) -> Dict[str, str]:
-        """Read-only route/tunnel status for the identity."""
-        raise NotImplementedError
-
-    def status(self, label: str,
-               run: Optional[RunFn] = None) -> Dict[str, str]:
-        """Read-only liveness: conf present, netns up, egress IP."""
-        raise NotImplementedError
-
-
-class WarpProvider(Provider):
-    """Cloudflare Warp provider on the established warp-* structures.
-
-    Identity: /etc/netvm/<label>.conf via netvm-new-identity.sh
-    (operator-authorized). Network: warp-<label> netns via
-    netvm-node-up.sh / netvm-node-down.sh. Exec: netvm-exec.sh.
-    Scopes are NOT nodes: no chrome-box profile, no NODES.md entry.
-    """
-
-    name = "warp"
-    ready = True
-
-    def _conf_exists(self, label: str, run: RunFn) -> bool:
-        rc, _ = run(["test", "-f", "/etc/netvm/%s.conf" % label],
-                    timeout=10)
-        return rc == 0
-
-    def provision(self, label: str,
-                  run: Optional[RunFn] = None) -> Dict[str, str]:
-        run = run or _run
-        check_label(label)
-        steps = []
-        if not self._conf_exists(label, run):
-            rc, out = run(["sudo", "-n", str(BIN_DIR / "netvm-new-identity.sh"),
-                           label], timeout=300)
-            if rc != 0:
-                raise ProviderError("warp identity failed for %s: %s"
-                                    % (label, out[-200:]))
-            steps.append("identity=new")
-        else:
-            steps.append("identity=exists")
-        rc, out = run(["sudo", "-n", str(BIN_DIR / "netvm-node-up.sh"),
-                       label], timeout=300)
-        if rc != 0:
-            raise ProviderError("netns up failed for %s: %s"
-                                % (label, out[-200:]))
-        steps.append("netns=up")
-        return {"ok": "true", "label": label, "netns": "warp-" + label,
-                "steps": ",".join(steps)}
-
-    def teardown(self, label: str,
-                 run: Optional[RunFn] = None) -> Dict[str, str]:
-        run = run or _run
-        check_label(label)
-        rc, out = run(["sudo", "-n", str(BIN_DIR / "netvm-node-down.sh"),
-                       label], timeout=120)
-        if rc != 0:
-            raise ProviderError("netns down failed for %s: %s"
-                                % (label, out[-200:]))
-        return {"ok": "true", "label": label, "netns": "warp-" + label}
-
-    def cycle(self, label: str,
-              run: Optional[RunFn] = None) -> Dict[str, str]:
-        """Fresh warp identity: down + remove conf + provision.
-
-        Conf removal needs root on /etc/netvm; when denied, the old
-        identity is left intact (netns down) and the operator gets the
-        exact human step instead of a half-rotated state.
-        """
-        run = run or _run
-        check_label(label)
-        self.teardown(label, run=run)
-        rc, out = run(["sudo", "-n", "rm", "-f",
-                       "/etc/netvm/%s.conf" % label], timeout=30)
-        if rc != 0:
-            raise ProviderError(
-                "rotation paused for %s: cannot remove old identity "
-                "(%s). Human: sudo rm /etc/netvm/%s.conf, then cycle "
-                "again." % (label, out[-120:], label))
-        return self.provision(label, run=run)
-
-    def exec(self, label: str, cmd: List[str],
-             run: Optional[RunFn] = None) -> Tuple[int, str]:
-        run = run or _run
-        check_label(label)
-        if not cmd:
-            raise ProviderError("exec needs a command")
-        return run([str(BIN_DIR / "netvm-exec.sh"), label, "--"] + cmd,
-                   timeout=120)
-
-    def routes(self, label: str,
-               run: Optional[RunFn] = None) -> Dict[str, str]:
-        run = run or _run
-        check_label(label)
-        netns = "warp-" + label
-        _, route_out = run(["sudo", "-n", "ip", "netns", "exec", netns,
-                            "ip", "route"], timeout=30)
-        _, wg_out = run(["sudo", "-n", "ip", "netns", "exec", netns,
-                         "wg", "show"], timeout=30)
-        return {"label": label, "netns": netns, "routes": route_out,
-                "wireguard": wg_out}
-
-    def status(self, label: str,
-               run: Optional[RunFn] = None) -> Dict[str, str]:
-        run = run or _run
-        check_label(label)
-        conf = self._conf_exists(label, run)
-        rc, out = run(["ip", "netns", "list"], timeout=10)
-        up = rc == 0 and ("warp-" + label) in out
-        egress = ""
-        if conf and up:
-            rc, eg = self.exec(label, ["curl", "-s", "--max-time", "8",
-                                       "https://api.ipify.org"], run=run)
-            egress = eg.strip().splitlines()[-1] if rc == 0 and eg.strip() \
-                else ""
-        return {"label": label, "conf": "yes" if conf else "no",
-                "netns": "up" if up else "down", "egress": egress or "n/a"}
-
-
-class GenericWireGuardProvider(Provider):
-    """BOILERPLATE: bring-your-own WireGuard confinement.
-
-    Intended structure: the operator supplies a wg conf out of band
-    (same root-0600 handling as Warp confs — never read here);
-    provision creates warp-<label> netns + veth/NAT exactly like
-    WarpProvider but consumes the supplied conf instead of a
-    Cloudflare-registered identity. Cycle swaps to the next supplied
-    conf. Implement when the first non-Cloudflare tunnel is needed.
-    """
-
-    name = "wireguard"
-
-
-class SocksProxyProvider(Provider):
-    """BOILERPLATE: per-scope SOCKS5 forward, no netns.
-
-    Intended structure: provision opens a dedicated local forward
-    (ssh -D style) per scope label and records its port; exec runs
-    commands with ALL_PROXY scoped to that port instead of entering a
-    netns; cycle re-establishes the forward via a fresh egress.
-    Implement when a scope needs proxy semantics without tunnels.
-    """
-
-    name = "socks"
-
-
-if __name__ == "__main__":
-    print("identity-provider.py is a library (see identity-broker.py)")
-    sys.exit(2)

bin/identity-resolve.py

@@ -1,186 +0,0 @@
-#!/usr/bin/env python3
-"""identity-resolve.py — Pure identity resolution for the identity plane.
-
-Reads identity-map.json (fingerprints only, never key material) and
-resolves an API-key fingerprint to its network-identity scope:
-
-  api_key -> account_origin(s); one origin rolls scope UP to the
-  umbrella account, two or more keep scope DOWN at the key itself.
-
-This module is pure + total (missing/corrupt map -> empty, unknown
-fingerprint -> None). CLI output carries emails and fingerprints only;
-key bytes never appear here — there is no code path that reads them
-except `fp`, which hashes stdin and prints only the digest.
-
-Usage:
-  identity-resolve.py fp < keyfile          # print sha256: fingerprint
-  identity-resolve.py lookup <fingerprint>  # print scope JSON
-  identity-resolve.py check                 # validate map schema
-"""
-
-from __future__ import annotations
-
-import argparse
-import hashlib
-import json
-import re
-import sys
-from pathlib import Path
-from typing import Any, Dict, List, Optional
-
-REPO_ROOT = Path(__file__).resolve().parent.parent
-MAP_FILE = REPO_ROOT / "identity-map.json"
-
-LABEL_RE = re.compile(r"^[a-z0-9][a-z0-9-]{0,22}$")
-
-
-def fingerprint_hex(material: bytes) -> str:
-    """sha256: fingerprint of raw key bytes."""
-    return "sha256:" + hashlib.sha256(material).hexdigest()
-
-
-def load_map(path: str | Path = MAP_FILE) -> Dict[str, Any]:
-    """Load the identity map. Missing/corrupt -> {"accounts": {}}."""
-    try:
-        with open(path, "r") as f:
-            data = json.load(f)
-        if isinstance(data, dict) and isinstance(
-                data.get("accounts"), dict):
-            return data
-    except Exception:
-        pass
-    return {"accounts": {}}
-
-
-def find_key(map_data: Dict[str, Any],
-             fp: str) -> Optional[Dict[str, Any]]:
-    """Locate a key record by fingerprint.
-
-    Returns {"email", "key"} or None. Top-level '_' entries ignored.
-    """
-    if not fp:
-        return None
-    accounts = map_data.get("accounts")
-    if not isinstance(accounts, dict):
-        return None
-    for email, rec in accounts.items():
-        if not isinstance(rec, dict):
-            continue
-        keys = rec.get("keys")
-        if not isinstance(keys, list):
-            continue
-        for k in keys:
-            if isinstance(k, dict) and k.get("fp") == fp:
-                return {"email": email, "key": k}
-    return None
-
-
-def resolve_scope(map_data: Dict[str, Any],
-                  fp: str) -> Optional[Dict[str, Any]]:
-    """Resolve a fingerprint to its scope unit.
-
-    Single origin -> {"scope": "account", "unit": email, ...}.
-    Multiple origins -> {"scope": "key", "unit": fp, ...}.
-    Unknown fingerprint -> None. Result carries emails + fingerprints
-    only (no key material exists anywhere in this module).
-    """
-    found = find_key(map_data, fp)
-    if found is None:
-        return None
-    key = found["key"]
-    origins = key.get("origins")
-    if not isinstance(origins, list) or not origins:
-        return None
-    origins = [str(o) for o in origins]
-    if len(origins) == 1:
-        return {"scope": "account", "unit": origins[0],
-                "email": found["email"], "origins": origins,
-                "label": key.get("label", "")}
-    return {"scope": "key", "unit": fp, "email": found["email"],
-            "origins": origins, "label": key.get("label", "")}
-
-
-def scope_slug(scope: Dict[str, Any]) -> str:
-    """Deterministic netvm label for a scope (fits label validation).
-
-    Account scopes: id-<email-fragment>-<hash7>. Key scopes:
-    id-k-<fp-hex-prefix>. Always matches ^[a-z0-9][a-z0-9-]{0,22}$.
-    """
-    unit = str(scope.get("unit", ""))
-    if scope.get("scope") == "key":
-        hexpart = re.sub(r"[^0-9a-f]", "", unit.lower())[:12] or "0"
-        return "id-k-%s" % hexpart
-    frag = re.sub(r"[^a-z0-9]+", "-", unit.lower()).strip("-")[:12]
-    frag = frag.strip("-") or "x"
-    tag = hashlib.sha256(unit.encode()).hexdigest()[:7]
-    return "id-%s-%s" % (frag, tag)
-
-
-def check_map(map_data: Dict[str, Any]) -> List[str]:
-    """Validate map schema. Returns a list of problem strings (empty OK)."""
-    problems: List[str] = []
-    accounts = map_data.get("accounts")
-    if not isinstance(accounts, dict):
-        return ["top-level 'accounts' must be an object"]
-    seen_fps: Dict[str, str] = {}
-    for email, rec in accounts.items():
-        if not isinstance(email, str) or "@" not in email:
-            problems.append("account key %r is not an email" % (email,))
-        if not isinstance(rec, dict) or not isinstance(
-                rec.get("keys"), list):
-            problems.append("account %r: 'keys' must be a list" % (email,))
-            continue
-        for i, k in enumerate(rec["keys"]):
-            where = "%s.keys[%d]" % (email, i)
-            if not isinstance(k, dict):
-                problems.append("%s: not an object" % where)
-                continue
-            fp = k.get("fp", "")
-            if not re.fullmatch(r"sha256:[0-9a-f]{64}", str(fp)):
-                problems.append("%s: bad fingerprint %r" % (where, fp))
-            elif fp in seen_fps:
-                problems.append("%s: fingerprint already listed under %s"
-                                % (where, seen_fps[fp]))
-            else:
-                seen_fps[fp] = email
-            origins = k.get("origins")
-            if not isinstance(origins, list) or not origins or not all(
-                    isinstance(o, str) and o for o in origins):
-                problems.append("%s: 'origins' must be a non-empty "
-                                "string list" % where)
-    return problems
-
-
-def main(argv: Optional[List[str]] = None) -> int:
-    ap = argparse.ArgumentParser(prog="identity-resolve.py")
-    ap.add_argument("--map", default=str(MAP_FILE),
-                    help="identity map (default: identity-map.json)")
-    sub = ap.add_subparsers(dest="cmd", required=True)
-    sub.add_parser("fp", help="print sha256: fingerprint of stdin bytes")
-    p = sub.add_parser("lookup", help="resolve a fingerprint to scope JSON")
-    p.add_argument("fp")
-    sub.add_parser("check", help="validate the map schema")
-    args = ap.parse_args(argv)
-    if args.cmd == "fp":
-        sys.stdout.write(fingerprint_hex(sys.stdin.buffer.read()) + "\n")
-        return 0
-    if args.cmd == "lookup":
-        scope = resolve_scope(load_map(args.map), args.fp)
-        if scope is None:
-            print("unknown fingerprint (not in map)")
-            return 1
-        scope["slug"] = scope_slug(scope)
-        print(json.dumps(scope, indent=2))
-        return 0
-    problems = check_map(load_map(args.map))
-    if problems:
-        print("%s INVALID:" % args.map)
-        for prob in problems:
-            print("  - %s" % prob)
-        return 1
-    print("%s OK" % args.map)
-    return 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/identity-variance-test.py

@@ -1,296 +0,0 @@
-#!/usr/bin/env python3
-"""Identity variance testing framework.
-
-Sends gentle, controlled probes from each NetVM node's Warp identity and
-measures per-identity response behavior so that future rate limits can be
-attributed (per-identity vs per-IP vs time-based vs random).
-
-Design goals:
-  - Very gentle traffic: 2 probes per node per cycle, default 10-min cycle.
-  - Never logs secrets: only sha256 hashes of WireGuard private keys.
-  - JSONL results log: one line per probe, machine-readable.
-
-Usage:
-  identity-variance-test.py probe    # run one probe cycle over all nodes
-  identity-variance-test.py analyze [--since HOURS] [--log PATH]
-
-Probes per node:
-  1. https://www.cloudflare.com/cdn-cgi/trace  (identity + egress IP Cloudflare sees)
-  2. https://muse.ai/                          (production-relevant landing page)
-
-Exit codes: 0 ok, 1 partial (some nodes failed), 2 fatal.
-"""
-
-import argparse
-import hashlib
-import json
-import os
-import re
-import subprocess
-import sys
-import time
-from datetime import datetime, timezone
-
-DEFAULT_LOG = os.path.expanduser(
-    "~/Projects/NetVM/.state/identity-variance/variance.jsonl"
-)
-NODES = ["muse", "pip", "646", "opm", "def"]
-
-PROBES = [
-    ("cf-trace", "https://www.cloudflare.com/cdn-cgi/trace"),
-    ("muse-landing", "https://muse.ai/"),
-]
-
-CURL_TIMEOUT = 15
-
-
-def sh(cmd, timeout=30):
-    """Run cmd (list) and return (rc, stdout, stderr)."""
-    try:
-        p = subprocess.run(
-            cmd, capture_output=True, text=True, timeout=timeout
-        )
-        return p.returncode, p.stdout, p.stderr
-    except subprocess.TimeoutExpired as e:
-        return 124, (e.stdout or ""), "timeout"
-    except FileNotFoundError:
-        return 127, "", "command not found"
-
-
-def node_netns_exists(node):
-    rc, out, _ = sh(["sudo", "-n", "ip", "netns", "list"])
-    return rc == 0 and f"warp-{node}" in out
-
-
-def identity_hash(node):
-    """Return truncated sha256 of the node's WireGuard private key (never the key)."""
-    rc, out, _ = sh(["sudo", "-n", "cat", f"/etc/netvm/{node}.conf"])
-    if rc == 0:
-        for line in out.splitlines():
-            m = re.match(r"\s*PrivateKey\s*=\s*(\S+)", line)
-            if m:
-                return hashlib.sha256(m.group(1).encode()).hexdigest()[:16]
-    return None
-
-
-def probe_node(node, target_name, url):
-    """One probe from inside the node's netns. Returns dict."""
-    # -w fields: http_code, time_total, size_download, remote_ip
-    fmt = "%{http_code} %{time_total} %{size_download} %{remote_ip}"
-    cmd = [
-        "sudo", "-n", "ip", "netns", "exec", f"warp-{node}",
-        "curl", "-s", "-o", "/dev/null", "-m", str(CURL_TIMEOUT),
-        "-w", fmt, url,
-    ]
-    started = datetime.now(timezone.utc)
-    rc, out, err = sh(cmd, timeout=CURL_TIMEOUT + 10)
-    ended = datetime.now(timezone.utc)
-
-    rec = {
-        "ts": started.isoformat(),
-        "node": node,
-        "probe": target_name,
-        "url": url,
-        "curl_rc": rc,
-    }
-    parts = out.strip().split()
-    if rc == 0 and len(parts) == 4:
-        try:
-            rec["http_status"] = int(parts[0])
-        except ValueError:
-            rec["http_status"] = None
-        try:
-            rec["response_ms"] = round(float(parts[1]) * 1000, 1)
-        except ValueError:
-            rec["response_ms"] = None
-        try:
-            rec["bytes"] = int(parts[2])
-        except ValueError:
-            rec["bytes"] = None
-        rec["remote_ip"] = parts[3] if parts[3] != "0.0.0.0" else None
-    else:
-        rec["http_status"] = None
-        rec["response_ms"] = None
-        rec["bytes"] = None
-        rec["remote_ip"] = None
-        rec["curl_error"] = (err or "curl failed").strip()[:200]
-
-    rec["rate_limited"] = rec["http_status"] in (429,)
-    rec["blocked"] = rec["http_status"] in (403,)
-    rec["ok"] = rec["http_status"] is not None and 200 <= rec["http_status"] < 400
-    rec["elapsed_wall_ms"] = round(
-        (ended - started).total_seconds() * 1000, 1
-    )
-    return rec
-
-
-def egress_ip(node):
-    """Best-effort egress IP as seen from inside the netns."""
-    rc, out, _ = sh([
-        "sudo", "-n", "ip", "netns", "exec", f"warp-{node}",
-        "curl", "-s", "-m", "10", "https://api.ipify.org",
-    ], timeout=20)
-    if rc == 0 and re.fullmatch(r"[0-9a-fA-F.:]+", out.strip()):
-        return out.strip()
-    return None
-
-
-def run_probe_cycle(log_path):
-    os.makedirs(os.path.dirname(log_path), exist_ok=True)
-    results = []
-    any_ok, any_fail = False, False
-    idhash = {}
-    for node in NODES:
-        if not node_netns_exists(node):
-            results.append({
-                "ts": datetime.now(timezone.utc).isoformat(),
-                "node": node, "probe": "node-skip",
-                "ok": False, "note": "netns warp-%s missing" % node,
-            })
-            any_fail = True
-            continue
-        idhash[node] = identity_hash(node)
-        for target_name, url in PROBES:
-            rec = probe_node(node, target_name, url)
-            rec["identity_hash"] = idhash[node]
-            results.append(rec)
-            if rec["ok"]:
-                any_ok = True
-            else:
-                any_fail = True
-            time.sleep(1)  # gentle pacing between probes
-
-    # Attach egress IP per node (one lookup per node, cached per cycle).
-    egress = {}
-    for node in {r["node"] for r in results if r.get("probe") != "node-skip"}:
-        egress[node] = egress_ip(node)
-    for rec in results:
-        if rec.get("probe") != "node-skip":
-            rec["egress_ip"] = egress.get(rec["node"])
-
-    with open(log_path, "a") as f:
-        for rec in results:
-            f.write(json.dumps(rec) + "\n")
-
-    print(json.dumps({
-        "cycle_ts": datetime.now(timezone.utc).isoformat(),
-        "log": log_path,
-        "records": len(results),
-        "ok": sum(1 for r in results if r.get("ok")),
-        "failed": sum(1 for r in results if not r.get("ok")),
-        "nodes": sorted({r["node"] for r in results}),
-        "egress": egress,
-        "identities": idhash,
-    }, indent=2))
-
-    if not any_ok:
-        return 2
-    return 1 if any_fail else 0
-
-
-def analyze(log_path, since_hours=None):
-    if not os.path.exists(log_path):
-        print("no log yet at %s" % log_path, file=sys.stderr)
-        return 2
-    cutoff = None
-    if since_hours:
-        cutoff = time.time() - since_hours * 3600
-
-    per_node = {}
-    rate_limit_events = []
-    identity_changes = {}
-    egress_by_node = {}
-
-    with open(log_path) as f:
-        for line in f:
-            line = line.strip()
-            if not line:
-                continue
-            try:
-                r = json.loads(line)
-            except json.JSONDecodeError:
-                continue
-            if r.get("probe") == "node-skip":
-                continue
-            try:
-                ts = datetime.fromisoformat(r["ts"]).timestamp()
-            except (ValueError, KeyError):
-                continue
-            if cutoff and ts < cutoff:
-                continue
-            node = r["node"]
-            st = per_node.setdefault(node, {
-                "probes": 0, "ok": 0, "ms": [], "429": 0, "403": 0,
-                "statuses": {}, "identities": set(), "probes_by_target": {},
-            })
-            st["probes"] += 1
-            if r.get("ok"):
-                st["ok"] += 1
-            if r.get("response_ms") is not None:
-                st["ms"].append(r["response_ms"])
-            if r.get("rate_limited"):
-                st["429"] += 1
-                rate_limit_events.append((r["ts"], node, r["probe"]))
-            if r.get("blocked"):
-                st["403"] += 1
-            s = r.get("http_status")
-            st["statuses"][str(s)] = st["statuses"].get(str(s), 0) + 1
-            if r.get("identity_hash"):
-                st["identities"].add(r["identity_hash"])
-            t = r.get("probe")
-            st["probes_by_target"][t] = st["probes_by_target"].get(t, 0) + 1
-            if r.get("egress_ip"):
-                egress_by_node.setdefault(node, set()).add(r["egress_ip"])
-
-    def pct(vals, p):
-        if not vals:
-            return None
-        s = sorted(vals)
-        return round(s[min(len(s) - 1, int(p / 100 * len(s)))], 1)
-
-    report = {"log": log_path, "nodes": {}}
-    for node in sorted(per_node):
-        st = per_node[node]
-        report["nodes"][node] = {
-            "probes": st["probes"],
-            "ok_rate": round(st["ok"] / st["probes"], 3) if st["probes"] else 0,
-            "latency_ms": {
-                "mean": round(sum(st["ms"]) / len(st["ms"]), 1) if st["ms"] else None,
-                "p50": pct(st["ms"], 50),
-                "p99": pct(st["ms"], 99),
-            },
-            "http_429": st["429"],
-            "http_403": st["403"],
-            "statuses": st["statuses"],
-            "identity_rotations": max(0, len(st["identities"]) - 1),
-            "egress_ips": sorted(egress_by_node.get(node, set())),
-        }
-        if len(st["identities"]) > 1:
-            identity_changes[node] = sorted(st["identities"])
-
-    # Divergence analysis: did all nodes see the same fate at the same time?
-    report["rate_limit_events"] = [
-        {"ts": ts, "node": n, "probe": p} for ts, n, p in rate_limit_events[-50:]
-    ]
-    report["identity_changes"] = identity_changes
-    print(json.dumps(report, indent=2))
-    return 0
-
-
-def main():
-    ap = argparse.ArgumentParser(description="Identity variance testing framework")
-    sub = ap.add_subparsers(dest="cmd", required=True)
-    p_probe = sub.add_parser("probe", help="run one probe cycle")
-    p_probe.add_argument("--log", default=DEFAULT_LOG)
-    p_an = sub.add_parser("analyze", help="summarize variance data")
-    p_an.add_argument("--log", default=DEFAULT_LOG)
-    p_an.add_argument("--since", type=float, default=None,
-                      help="only include last N hours")
-    args = ap.parse_args()
-    if args.cmd == "probe":
-        sys.exit(run_probe_cycle(args.log))
-    sys.exit(analyze(args.log, args.since))
-
-
-if __name__ == "__main__":
-    main()

bin/invite.py

@@ -1,276 +0,0 @@
-#!/usr/bin/env python3
-"""invite.py — Muse.ai invite codes + usage via the agent browsers.
-
-Find: GET /api/hatch/invite via in-page fetch (primary; proven live),
-      Invite-button popover parse (DOM fallback).
-Redeem: POST /api/hatch/invite-code/redeem via in-page fetch.
-Usage: Settings > General DOM read (weekly reset, % used, additional
-      tokens) through the hatch_menu tree (dialog/mouse/tabs);
-      settings-nav primitives live there, this module keeps the
-      invite API + popover flows and the flat CLI parse.
-
-Server eligibility (observed live): one redemption per account
-(already_redeemed); 48h window from joining (window_expired); codes
-carry limited uses (used_up) and can be revoked. Inviter rewards
-accrue regardless of the inviter's own redemption state.
-
-Caller errors (unknown node, malformed code) raise InviteError before
-any CDP traffic. Transport/CDP failures return {"ok": False, ...}.
-"""
-
-import re
-import time
-
-from approvals import VALID_NODES, cdp_evaluate, get_cdp_ws
-from hatch_menu import dialog as menu_dialog
-from hatch_menu.mouse import close as _close
-from hatch_menu.mouse import escape as _escape
-from hatch_menu.tabs import general as general_tab
-
-
-class InviteError(ValueError):
-    """Caller error: unknown node or malformed code. Raised before CDP."""
-
-
-CODE_RE = re.compile(r"^[A-Z0-9]{6}$")
-CODE_REVEALED_RE = re.compile(r"Invite code revealed:\s*([A-Z0-9]{6})")
-
-JS_INVITE_GET = """(async () => {
-  try {
-    const r = await fetch('/api/hatch/invite',
-      {method: 'GET', cache: 'no-store'});
-    return {status: r.status, data: await r.json()};
-  } catch (e) { return {error: String(e).slice(0, 200)}; }
-})()"""
-
-# %s is a validated [A-Z0-9]{6} code: injection-safe by construction.
-JS_REDEEM_TMPL = """(async () => {
-  try {
-    const r = await fetch('/api/hatch/invite-code/redeem', {method: 'POST',
-      headers: {'Content-Type': 'application/json'},
-      body: JSON.stringify({code: '%s', supportsRedemptionStatus: true})});
-    return {status: r.status, ok: r.ok, data: await r.json()};
-  } catch (e) { return {error: String(e).slice(0, 200)}; }
-})()"""
-
-JS_INVITE_CLICK = ("(() => { const b = document.querySelector("
-                   "'[data-testid=\"hatch-invite-friends-button\"]');"
-                   " if (!b) return 'NO_BUTTON'; b.click();"
-                   " return 'CLICKED'; })()")
-
-JS_POPOVER_TEXT = ("(() => { const d = document.querySelector("
-                   "'[data-slot=\"popover-content\"]');"
-                   " return d ? d.innerText : null; })()")
-
-def normalize_code(code):
-    """Uppercase/strip a code; None unless 6-char A-Z0-9."""
-    if not isinstance(code, str):
-        return None
-    c = code.strip().upper()
-    return c if CODE_RE.fullmatch(c) else None
-
-
-def _check_node(node):
-    if node not in VALID_NODES:
-        raise InviteError("unknown node: %r (valid: %s)"
-                          % (node, ", ".join(VALID_NODES)))
-
-
-def open_settings(ws):
-    """Open the Settings dialog via the dock menu. True when open.
-
-    Public shim over hatch_menu.dialog.open_settings (single copy).
-    """
-    return menu_dialog.open_settings(ws)
-
-
-def click_settings_tab(ws, name):
-    """Open a Settings dialog tab by visible name. True when open.
-
-    Public shim over hatch_menu.dialog.goto_tab (single copy).
-    """
-    return menu_dialog.goto_tab(ws, name)
-
-
-def _popover_code(ws):
-    """Invite code via the main-chat popover (DOM fallback). None if absent."""
-    try:
-        if cdp_evaluate(ws, JS_INVITE_CLICK, timeout=5.0) != "CLICKED":
-            return None
-        time.sleep(1.0)
-        text = cdp_evaluate(ws, JS_POPOVER_TEXT, timeout=5.0) or ""
-    except Exception:
-        return None
-    finally:
-        _escape(ws)
-    m = CODE_REVEALED_RE.search(text)
-    return m.group(1) if m else None
-
-
-def get_invite(node, timeout=8.0):
-    """Per-agent invite state. API first, popover fallback for the code."""
-    _check_node(node)
-    try:
-        ws, _ = get_cdp_ws(node)
-    except Exception as e:
-        return {"ok": False, "node": node,
-                "error": "CDP unreachable: %s" % e}
-    try:
-        try:
-            res = cdp_evaluate(ws, JS_INVITE_GET, await_promise=True,
-                               timeout=timeout)
-        except Exception as e:
-            res = {"error": "%s: %s" % (type(e).__name__, e)}
-        if isinstance(res, dict) and res.get("status") == 200 \
-                and isinstance(res.get("data"), dict) \
-                and res["data"].get("code"):
-            d = res["data"]
-            return {"ok": True, "node": node, "source": "api",
-                    "code": d.get("code"),
-                    "has_redeemed": d.get("has_redeemed_invite_code"),
-                    "uses_remaining": d.get("uses_remaining"),
-                    "use_count": d.get("use_count"),
-                    "reward": d.get("reward")}
-        code = _popover_code(ws)
-        if code:
-            return {"ok": True, "node": node, "source": "dom",
-                    "code": code, "has_redeemed": None,
-                    "uses_remaining": None, "use_count": None,
-                    "reward": None}
-        if isinstance(res, dict):
-            detail = res.get("error", "invite API failed")
-        else:
-            detail = "invite API failed"
-        return {"ok": False, "node": node,
-                "error": "invite lookup failed (%s); popover has no code"
-                % detail}
-    finally:
-        _close(ws)
-
-
-def redeem_invite(node, code, timeout=12.0):
-    """Redeem CODE on node. Returns ok / server reason + detail."""
-    c = normalize_code(code)
-    if c is None:
-        raise InviteError("malformed invite code: %r (want 6 chars A-Z0-9)"
-                          % (code,))
-    _check_node(node)
-    js = JS_REDEEM_TMPL % c
-    try:
-        ws, _ = get_cdp_ws(node)
-    except Exception as e:
-        return {"ok": False, "node": node, "code": c,
-                "error": "CDP unreachable: %s" % e}
-    try:
-        try:
-            res = cdp_evaluate(ws, js, await_promise=True, timeout=timeout)
-        except Exception as e:
-            return {"ok": False, "node": node, "code": c,
-                    "error": "%s: %s" % (type(e).__name__, e)}
-        if not isinstance(res, dict) or "data" not in res:
-            if isinstance(res, dict):
-                detail = res.get("error", "empty redeem response")
-            else:
-                detail = "empty redeem response"
-            return {"ok": False, "node": node, "code": c,
-                    "reason": "unknown", "detail": detail}
-        data = res.get("data") or {}
-        if res.get("ok") and data.get("success"):
-            return {"ok": True, "node": node, "code": c,
-                    "redemption_status": data.get("redemptionStatus"),
-                    "detail": data.get("detail")}
-        return {"ok": False, "node": node, "code": c,
-                "reason": data.get("reason", "unknown"),
-                "detail": data.get("detail")}
-    finally:
-        _close(ws)
-
-
-def parse_usage_text(text):
-    """Flat usage parse (CLI contract) over the tree's General parser."""
-    nested = general_tab.parse_usage_text(text)
-    wk = (nested.get("weekly") if nested else None) or {}
-    add = (nested.get("additional") if nested else None) or {}
-    tokens = add.get("tokens_left")
-    return {
-        "weekly_reset": wk.get("resets_on"),
-        "weekly_used_pct": wk.get("pct_used"),
-        "additional_expires": "Never expires"
-        if add.get("never_expires") else None,
-        "additional_used_pct": add.get("pct_used"),
-        "additional_left": ("%s tokens left" % tokens)
-        if tokens else None,
-        "has_redeemed": bool(add.get("never_expires") or tokens
-                             or "additional tokens" in (text or "").lower()),
-    }
-
-
-def get_usage(node, timeout=8.0):
-    """Usage limits for one agent via Settings General (DOM read)."""
-    _check_node(node)
-    try:
-        ws, _ = get_cdp_ws(node)
-    except Exception as e:
-        return {"ok": False, "node": node,
-                "error": "CDP unreachable: %s" % e}
-    try:
-        if not open_settings(ws):
-            return {"ok": False, "node": node,
-                    "error": "settings dialog did not open"}
-        if not menu_dialog.goto_tab(ws, "General"):
-            return {"ok": False, "node": node,
-                    "error": "General tab did not open"}
-        # Note: Usage stats and redeem field can load asynchronously in the React/Radix tree.
-        # Poll up to timeout seconds for usage or redeem markers.
-        text = None
-        deadline = time.time() + timeout
-        while time.time() < deadline:
-            t = menu_dialog.dialog_text(ws)
-            if t and ("Weekly limit" in t or "Additional tokens" in t or "Redeem invite code" in t or "tokens left" in t or "% used" in t):
-                text = t
-                break
-            time.sleep(0.4)
-        if not text:
-            text = menu_dialog.dialog_text(ws)
-
-        if not text:
-            return {"ok": False, "node": node,
-                    "error": "empty settings dialog"}
-        out = {"ok": True, "node": node}
-        parsed = parse_usage_text(text)
-        stats_loaded = bool(
-            parsed.get("weekly_reset")
-            or parsed.get("weekly_used_pct") is not None
-            or parsed.get("has_redeemed")
-            or parsed.get("additional_left")
-        )
-        out.update(parsed)
-        out["stats_loaded"] = stats_loaded
-        if not stats_loaded:
-            out["note"] = "Usage stats did not render in Settings dialog"
-        return out
-    finally:
-        _escape(ws)
-        _close(ws)
-
-
-def fleet_invite_status(nodes=None):
-    """Invite state per node. Never raises; per-node error dicts."""
-    out = {}
-    for n in (nodes or list(VALID_NODES)):
-        try:
-            out[n] = get_invite(n)
-        except InviteError as e:
-            out[n] = {"ok": False, "node": n, "error": str(e)}
-    return out
-
-
-def fleet_usage(nodes=None):
-    """Usage limits per node. Never raises; per-node error dicts."""
-    out = {}
-    for n in (nodes or list(VALID_NODES)):
-        try:
-            out[n] = get_usage(n)
-        except InviteError as e:
-            out[n] = {"ok": False, "node": n, "error": str(e)}
-    return out

bin/invite_handler.py

@@ -1,790 +0,0 @@
-#!/usr/bin/env python3
-"""invite_handler.py: Invite code discovery, inspection, and redemption handler.
-
-Supports:
-  - Extracting agent invite codes via Main Chat DOM RPA & in-session API
-  - Redeeming invite codes via Settings Menu RPA & redemption endpoint
-  - Fleet-wide invite inventory, usage tracking, and agent salvage flows
-"""
-
-from __future__ import annotations
-
-import argparse
-import json
-import re
-import sys
-import time
-from dataclasses import asdict, dataclass
-from typing import Any, Dict, List, Optional
-
-try:
-    from approvals import cdp_evaluate, get_cdp_ws, get_node_pages
-    from settings_rpa import SettingsRPA, cdp_click_element_by_selector, cdp_send_escape
-except ImportError:
-    import os
-    sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
-    from approvals import cdp_evaluate, get_cdp_ws, get_node_pages
-    from settings_rpa import SettingsRPA, cdp_click_element_by_selector, cdp_send_escape
-
-
-VALID_NODES = ["muse", "pip", "646", "opm", "def", "dev"]
-
-
-@dataclass
-class InviteCodeInfo:
-    node: str
-    code: str
-    uses_remaining: int
-    use_count: int
-    max_uses: int
-    has_redeemed: bool
-    invite_state: str
-    reward: Optional[Dict[str, Any]]
-    method_used: str
-
-    def to_dict(self) -> Dict[str, Any]:
-        return asdict(self)
-
-
-@dataclass
-class RedemptionResult:
-    target_node: str
-    code: str
-    success: bool
-    status: str
-    reason: Optional[str]
-    detail: Optional[str]
-    method_used: str
-    field_missing: bool = False
-    loopback_notified: bool = False
-    loopback_detail: Optional[str] = None
-
-    def to_dict(self) -> Dict[str, Any]:
-        return asdict(self)
-
-
-def send_loopback_notice(
-    recipient: str,
-    target: str,
-    message: str,
-    sender: str = "super",
-    timeout: float = 15.0,
-) -> bool:
-    """Send a loopback notification DM via bin/dm.py to a sidechat without blocking or failing."""
-    try:
-        from pathlib import Path
-        bin_dir = Path(__file__).resolve().parent
-        dm_script = bin_dir / "dm.py"
-        if not dm_script.exists():
-            return False
-        import subprocess
-        cmd = [
-            sys.executable,
-            str(dm_script),
-            "send",
-            "--agent",
-            sender,
-            "--to",
-            recipient,
-            "--target",
-            target,
-            message,
-        ]
-        if target == "main":
-            cmd.insert(-1, "--allow-main-chat")
-        res = subprocess.run(cmd, capture_output=True, text=True, timeout=timeout)
-        return res.returncode == 0
-    except Exception:
-        return False
-
-
-class InviteHandler:
-    """Handles invite codes discovery, inspection, and redemption for NetVM nodes."""
-
-    def __init__(self, node: str, timeout: float = 4.0):
-        self.node = node
-        self.timeout = timeout
-        self.ws = None
-
-    def _dispatch_loopback(self, res: RedemptionResult, agent: str, target: str) -> bool:
-        """Post a loopback notification DM via box dm."""
-        msg = f"[BOX-INVITE-LOOPBACK] Node @{self.node} redeem status: {res.status}. Reason: {res.reason or 'unknown'}. Detail: {res.detail or '-'}"
-        ok = send_loopback_notice(recipient=agent, target=target, message=msg)
-        res.loopback_notified = ok
-        res.loopback_detail = f"Notified @{agent}/{target}" if ok else f"Failed to notify @{agent}/{target}"
-        return ok
-
-
-    def connect(self) -> InviteHandler:
-        if self.ws is None:
-            self.ws, _ = get_cdp_ws(self.node, timeout=self.timeout)
-        return self
-
-    def close(self) -> None:
-        if self.ws is not None:
-            try:
-                self.ws.close()
-            except Exception:
-                pass
-            self.ws = None
-
-    def __enter__(self) -> InviteHandler:
-        return self.connect()
-
-    def __exit__(self, exc_type, exc_val, exc_tb) -> None:
-        self.close()
-
-    def find_code_api(self) -> InviteCodeInfo:
-        """Fetch invite code and metadata directly via in-session /api/hatch/invite query."""
-        self.connect()
-        js_query = """(async () => {
-            try {
-                const resp = await fetch('/api/hatch/invite', {method: 'GET', cache: 'no-store'});
-                if (!resp.ok) return {error: `HTTP ${resp.status}`};
-                return await resp.json();
-            } catch(e) {
-                return {error: e.toString()};
-            }
-        })()"""
-        res = cdp_evaluate(self.ws, js_query, await_promise=True, timeout=self.timeout)
-        if not res or "error" in res or "code" not in res:
-            err = res.get("error", "Invalid response") if res else "No response"
-            raise RuntimeError(f"Failed to fetch invite info for {self.node}: {err}")
-
-        return InviteCodeInfo(
-            node=self.node,
-            code=res.get("code", ""),
-            uses_remaining=res.get("uses_remaining", 0),
-            use_count=res.get("use_count", 0),
-            max_uses=res.get("max_uses", 30),
-            has_redeemed=bool(res.get("has_redeemed_invite_code", False)),
-            invite_state=res.get("invite_state", "UNKNOWN"),
-            reward=res.get("reward"),
-            method_used="api",
-        )
-
-    def find_code_dom(self) -> InviteCodeInfo:
-        """Find invite code by navigating to Main Chat and opening the Invite popover."""
-        self.connect()
-        # 1. Ensure main chat home is active
-        with SettingsRPA(self.node, timeout=self.timeout) as rpa:
-            rpa.ensure_main_chat()
-
-        time.sleep(0.3)
-        # Dismiss any open popovers first
-        cdp_send_escape(self.ws)
-        time.sleep(0.2)
-
-        # 2. Click Invite button in main chat
-        clicked = cdp_click_element_by_selector(self.ws, '[data-testid="hatch-invite-friends-button"]', timeout=self.timeout)
-        if not clicked:
-            # Fallback to button by text
-            js_click = """(() => {
-                const btn = Array.from(document.querySelectorAll('button')).find(b => (b.innerText||'').trim() === 'Invite');
-                if (btn) { btn.click(); return true; }
-                return false;
-            })()"""
-            clicked = cdp_evaluate(self.ws, js_click)
-
-        if not clicked:
-            raise RuntimeError(f"Could not click Invite button on {self.node}")
-
-        time.sleep(0.8)
-
-        # 3. Read popover contents
-        js_read_popover = """(() => {
-            const pop = document.querySelector('[data-slot="popover-content"]');
-            if (!pop) return {found: false};
-
-            // Look for revealed code span
-            const codeSpan = pop.querySelector('[data-pel-impression="invite_code_revealed_impression"]');
-            let code = codeSpan ? codeSpan.innerText.trim() : null;
-
-            // Regex fallback on text
-            const fullText = pop.innerText || '';
-            if (!code) {
-                const m = fullText.match(/Invite code revealed:\\s*([A-Z0-9]{6})/);
-                if (m) code = m[1];
-            }
-            if (!code) {
-                const m2 = fullText.match(/\\b([A-Z0-9]{6})\\b/);
-                if (m2) code = m2[1];
-            }
-
-            let usesLeft = 30;
-            const mUses = fullText.match(/(\\d+)\\s+uses\\s+left/);
-            if (mUses) usesLeft = parseInt(mUses[1], 10);
-
-            return {
-                found: true,
-                code: code,
-                uses_left: usesLeft,
-                text: fullText
-            };
-        })()"""
-        pop_res = cdp_evaluate(self.ws, js_read_popover)
-        cdp_send_escape(self.ws)
-
-        if not pop_res or not pop_res.get("found") or not pop_res.get("code"):
-            # Fall back to API if popover extraction failed
-            return self.find_code_api()
-
-        code = pop_res.get("code")
-        uses_left = pop_res.get("uses_left", 30)
-
-        # Query API for additional metadata
-        try:
-            api_info = self.find_code_api()
-            api_info.method_used = "dom+api"
-            return api_info
-        except Exception:
-            return InviteCodeInfo(
-                node=self.node,
-                code=code,
-                uses_remaining=uses_left,
-                use_count=30 - uses_left,
-                max_uses=30,
-                has_redeemed=True,
-                invite_state="ELIGIBLE",
-                reward=None,
-                method_used="dom",
-            )
-
-    def find_code(self, method: str = "auto") -> InviteCodeInfo:
-        """Find invite code using specified method ('auto', 'api', or 'dom')."""
-        if method == "api":
-            return self.find_code_api()
-        elif method == "dom":
-            return self.find_code_dom()
-        else:
-            # Auto: Try API first, fallback to DOM
-            try:
-                return self.find_code_api()
-            except Exception:
-                return self.find_code_dom()
-
-    def redeem_code_api(self, code: str) -> RedemptionResult:
-        """Redeem invite code using in-session /api/hatch/invite-code/redeem endpoint."""
-        self.connect()
-        clean_code = code.strip().upper()
-        js_redeem = f"""(async () => {{
-            try {{
-                const resp = await fetch('/api/hatch/invite-code/redeem', {{
-                    method: 'POST',
-                    headers: {{'Content-Type': 'application/json'}},
-                    body: JSON.stringify({{code: {json.dumps(clean_code)}, supportsRedemptionStatus: true}})
-                }});
-                const data = await resp.json();
-                return {{status_code: resp.status, ok: resp.ok, data: data}};
-            }} catch(e) {{
-                return {{error: e.toString()}};
-            }}
-        }})()"""
-        res = cdp_evaluate(self.ws, js_redeem, await_promise=True, timeout=self.timeout)
-        if not res or "error" in res:
-            err = res.get("error", "No response") if res else "Timeout"
-            return RedemptionResult(
-                target_node=self.node,
-                code=clean_code,
-                success=False,
-                status="error",
-                reason="transport_failure",
-                detail=err,
-                method_used="api",
-            )
-
-        data = res.get("data", {})
-        ok = res.get("ok", False) and data.get("success", False)
-        status = data.get("redemptionStatus") or ("redeemed" if ok else "failed")
-        reason = data.get("reason")
-        detail = data.get("detail")
-
-        return RedemptionResult(
-            target_node=self.node,
-            code=clean_code,
-            success=ok,
-            status=status,
-            reason=reason,
-            detail=detail,
-            method_used="api",
-        )
-
-    def redeem_code_dom(
-        self,
-        code: str,
-        notify_target: Optional[str] = None,
-        notify_agent: Optional[str] = None,
-    ) -> RedemptionResult:
-        """Redeem invite code using Settings Menu RPA -> Settings -> Redeem Invite Code dialog.
-        Gracefully handles missing entrypoint rows, missing input fields, and async loading races.
-        Falls back to in-page API automatically if DOM fields are absent, and dispatches DM loopback if requested."""
-        self.connect()
-        clean_code = code.strip().upper()
-        with SettingsRPA(self.node, timeout=self.timeout) as rpa:
-            # 1. Open Settings dialog
-            opened = rpa.open_settings_dialog()
-            if not opened:
-                # If dialog failed to open, try API fallback directly
-                api_res = self.redeem_code_api(clean_code)
-                if api_res.success:
-                    api_res.detail = f"Settings dialog could not open; redemption completed via API fallback ({api_res.detail or ''})".strip()
-                    return api_res
-
-                res = RedemptionResult(
-                    target_node=self.node,
-                    code=clean_code,
-                    success=False,
-                    status="error",
-                    reason="settings_open_failed",
-                    detail=f"Could not open Settings dialog via RPA on @{self.node}. API fallback: {api_res.detail or api_res.reason or 'failed'}",
-                    method_used="dom",
-                    field_missing=True,
-                )
-                if notify_target and notify_agent:
-                    self._dispatch_loopback(res, notify_agent, notify_target)
-                return res
-
-            rpa.select_tab("General")
-            time.sleep(0.3)
-
-            # 2. Check for "Redeem invite code" entry (poll up to 2.5s for React rendering)
-            js_find_and_click = """(() => {
-                const dialog = document.querySelector('[role="dialog"]');
-                if (!dialog) return {found: false, dialog_present: false};
-                const items = Array.from(dialog.querySelectorAll('button, div, span'));
-                const redeemItem = items.find(el => (el.innerText || '').trim() === 'Redeem invite code');
-                if (redeemItem) {
-                    redeemItem.click();
-                    return {found: true, dialog_present: true};
-                }
-                return {
-                    found: false,
-                    dialog_present: true,
-                    text: dialog.innerText || '',
-                    has_additional: (dialog.innerText || '').includes('Additional tokens')
-                };
-            })()"""
-            click_res = None
-            deadline = time.time() + 2.5
-            while time.time() < deadline:
-                click_res = cdp_evaluate(self.ws, js_find_and_click)
-                if click_res and click_res.get("found"):
-                    break
-                time.sleep(0.3)
-
-            if not click_res or not click_res.get("found"):
-                # "Redeem invite code" field is missing in General settings!
-                rpa.close_settings_dialog()
-
-                # Attempt automatic API fallback first
-                api_res = self.redeem_code_api(clean_code)
-                if api_res.success:
-                    api_res.detail = f"Redeem field was missing in Settings DOM; redeemed successfully via API fallback! ({api_res.detail or ''})".strip()
-                    return api_res
-
-                # Diagnose why field is missing
-                diag_text = click_res.get("text", "") if click_res else ""
-                has_extra = click_res.get("has_additional", False) if click_res else False
-
-                is_already = has_extra or "Additional tokens" in diag_text
-                if not is_already:
-                    try:
-                        api_check = self.find_code_api()
-                        is_already = bool(api_check.has_redeemed)
-                    except Exception:
-                        pass
-
-                if is_already:
-                    status = "already_redeemed"
-                    reason = "already_redeemed"
-                    detail = f"Node @{self.node} has already redeemed an invite code (entrypoint hidden by active Additional tokens ticker)."
-                elif api_res.reason in ["window_expired", "already_redeemed", "invalid", "used_up"]:
-                    status = api_res.status
-                    reason = api_res.reason
-                    detail = f"Redeem invite code field not present in Settings on @{self.node}: server reports {api_res.reason} ({api_res.detail or ''})."
-                else:
-                    status = "entrypoint_not_found"
-                    reason = "field_missing"
-                    detail = f"Redeem invite code field missing in General settings on @{self.node} (account may be past 48h onboarding window or already redeemed)."
-
-                res = RedemptionResult(
-                    target_node=self.node,
-                    code=clean_code,
-                    success=False,
-                    status=status,
-                    reason=reason,
-                    detail=detail,
-                    method_used="dom",
-                    field_missing=True,
-                )
-                if notify_target and notify_agent:
-                    self._dispatch_loopback(res, notify_agent, notify_target)
-                return res
-
-            # 3. Handle HatchInviteRedemptionDialog (sub-dialog opened by clicking Redeem)
-            # Poll up to 2.5s for input box to mount
-            js_input_code = f"""(() => {{
-                // Look for dialog titled "Redeem a code" or input with label
-                const input = document.querySelector('input[aria-label="Invite code"]') ||
-                              document.querySelector('[role="dialog"] input[type="text"]');
-                if (!input) return {{found_input: false}};
-
-                input.focus();
-                input.value = {json.dumps(clean_code)};
-                input.dispatchEvent(new Event('input', {{bubbles: true}}));
-                input.dispatchEvent(new Event('change', {{bubbles: true}}));
-
-                input.dispatchEvent(new KeyboardEvent('keydown', {{key: 'Enter', code: 'Enter', keyCode: 13, which: 13, bubbles: true}}));
-                return {{found_input: true}};
-            }})()"""
-            input_res = None
-            deadline = time.time() + 2.5
-            while time.time() < deadline:
-                input_res = cdp_evaluate(self.ws, js_input_code)
-                if input_res and input_res.get("found_input"):
-                    break
-                time.sleep(0.3)
-
-            if not input_res or not input_res.get("found_input"):
-                # Subdialog opened or clicked, but input box is absent!
-                cdp_send_escape(self.ws)
-                rpa.close_settings_dialog()
-
-                # Automatic API fallback
-                api_res = self.redeem_code_api(clean_code)
-                if api_res.success:
-                    api_res.detail = f"Redemption input box was missing in dialog; redeemed successfully via API fallback! ({api_res.detail or ''})".strip()
-                    return api_res
-
-                res = RedemptionResult(
-                    target_node=self.node,
-                    code=clean_code,
-                    success=False,
-                    status=api_res.status or "dom_input_missing",
-                    reason=api_res.reason or "input_field_missing",
-                    detail=f"Invite code input field was not found in redemption dialog on @{self.node}. API fallback: {api_res.detail or api_res.reason or 'failed'}.",
-                    method_used="dom",
-                    field_missing=True,
-                )
-                if notify_target and notify_agent:
-                    self._dispatch_loopback(res, notify_agent, notify_target)
-                return res
-
-            # Input submitted; poll for outcome text
-            time.sleep(1.2)
-            js_check_outcome = """(() => {
-                const dialog = document.querySelector('[role="dialog"]');
-                if (!dialog) return {found: false};
-                const text = dialog.innerText || '';
-
-                if (text.includes('Congratulations') || text.includes('successful')) {
-                    return {success: true, status: 'redeemed', detail: text};
-                }
-                if (text.includes('already redeemed')) {
-                    return {success: false, status: 'already_redeemed', detail: text};
-                }
-                if (text.includes('no longer valid') || text.includes('invalid')) {
-                    return {success: false, status: 'invalid', detail: text};
-                }
-                if (text.includes('used up')) {
-                    return {success: false, status: 'used_up', detail: text};
-                }
-                if (text.includes('passed') || text.includes('expired')) {
-                    return {success: false, status: 'window_expired', detail: text};
-                }
-                return {status: 'unknown', detail: text};
-            })()"""
-            outcome = None
-            deadline = time.time() + 2.5
-            while time.time() < deadline:
-                outcome = cdp_evaluate(self.ws, js_check_outcome)
-                if outcome and outcome.get("status") != "unknown":
-                    break
-                time.sleep(0.3)
-
-            rpa.close_settings_dialog()
-
-            if outcome and outcome.get("success"):
-                return RedemptionResult(
-                    target_node=self.node,
-                    code=clean_code,
-                    success=True,
-                    status="redeemed",
-                    reason=None,
-                    detail="Redemption successful via Settings RPA",
-                    method_used="dom",
-                )
-            elif outcome and outcome.get("status") in ["already_redeemed", "invalid", "used_up", "window_expired"]:
-                res = RedemptionResult(
-                    target_node=self.node,
-                    code=clean_code,
-                    success=False,
-                    status=outcome.get("status"),
-                    reason=outcome.get("status"),
-                    detail=outcome.get("detail"),
-                    method_used="dom",
-                )
-                if notify_target and notify_agent:
-                    self._dispatch_loopback(res, notify_agent, notify_target)
-                return res
-
-            # Fallback to API check if DOM did not confirm status
-            api_res = self.redeem_code_api(clean_code)
-            if not api_res.success and notify_target and notify_agent:
-                self._dispatch_loopback(api_res, notify_agent, notify_target)
-            return api_res
-
-    def redeem_code(
-        self,
-        code: str,
-        method: str = "auto",
-        notify_target: Optional[str] = None,
-        notify_agent: Optional[str] = None,
-    ) -> RedemptionResult:
-        """Redeem invite code using specified method ('auto', 'dom', or 'api')."""
-        if method == "dom":
-            return self.redeem_code_dom(code, notify_target=notify_target, notify_agent=notify_agent)
-        elif method == "api":
-            res = self.redeem_code_api(code)
-            if not res.success and notify_target and notify_agent:
-                self._dispatch_loopback(res, notify_agent, notify_target)
-            return res
-        else:
-            # Auto: Validate and attempt via API for reliability. Fallback to DOM on transport error.
-            res = self.redeem_code_api(code)
-            if not res.success and res.reason == "transport_failure":
-                res = self.redeem_code_dom(code, notify_target=notify_target, notify_agent=notify_agent)
-            elif not res.success and notify_target and notify_agent:
-                self._dispatch_loopback(res, notify_agent, notify_target)
-            return res
-
-
-
-def scan_fleet_invites(nodes: Optional[List[str]] = None) -> List[Dict[str, Any]]:
-    """Scan fleet nodes and return invite code details for each active node."""
-    target_nodes = nodes or ["muse", "pip", "646", "opm"]
-    results = []
-    for node in target_nodes:
-        try:
-            handler = InviteHandler(node)
-            with handler:
-                info = handler.find_code(method="auto")
-                results.append(info.to_dict())
-        except Exception as e:
-            results.append({
-                "node": node,
-                "code": None,
-                "uses_remaining": 0,
-                "use_count": 0,
-                "max_uses": 30,
-                "has_redeemed": None,
-                "invite_state": "UNREACHABLE",
-                "reward": None,
-                "method_used": "error",
-                "error": str(e),
-            })
-    return results
-
-
-def scan_fleet_usage(nodes: Optional[List[str]] = None) -> List[Dict[str, Any]]:
-    """Scan fleet nodes and return token usage for each active node."""
-    target_nodes = nodes or ["muse", "pip", "646", "opm"]
-    results = []
-    for node in target_nodes:
-        try:
-            with SettingsRPA(node) as rpa:
-                usage = rpa.read_usage()
-                results.append(usage.to_dict())
-        except Exception as e:
-            results.append({
-                "node": node,
-                "plan": "Unknown",
-                "weekly_reset_text": "Unreachable",
-                "weekly_percent_used": 0,
-                "extra_tokens_status": "Unknown",
-                "extra_percent_used": 0,
-                "extra_tokens_remaining": "Unknown",
-                "is_blocked": False,
-                "bars": [],
-                "error": str(e),
-            })
-    return results
-
-
-def salvage_blocked_node(
-    blocked_node: str = "646",
-    helper_node: Optional[str] = None,
-    notify_target: Optional[str] = None,
-    notify_agent: Optional[str] = None,
-) -> Dict[str, Any]:
-    """Salvage an out-of-tokens node by identifying its code and redeeming it on an eligible peer."""
-    # 1. Fetch blocked node invite code
-    with InviteHandler(blocked_node) as h_blocked:
-        blocked_info = h_blocked.find_code()
-
-    code_to_redeem = blocked_info.code
-    if not code_to_redeem:
-        return {
-            "success": False,
-            "blocked_node": blocked_node,
-            "error": f"Could not find invite code for blocked node {blocked_node}",
-        }
-
-    # 2. Check candidate helper nodes
-    candidates = [helper_node] if helper_node else [n for n in VALID_NODES if n != blocked_node]
-    eligible_peer = None
-
-    for peer in candidates:
-        try:
-            with InviteHandler(peer) as h_peer:
-                peer_info = h_peer.find_code()
-                if not peer_info.has_redeemed:
-                    eligible_peer = peer
-                    break
-        except Exception:
-            continue
-
-    if not eligible_peer:
-        res = {
-            "success": False,
-            "blocked_node": blocked_node,
-            "invite_code": code_to_redeem,
-            "error": "No existing fleet peer is currently eligible (all active peers have already redeemed an invite code). An onboarding agent or fresh client profile must redeem this code.",
-            "code_to_redeem": code_to_redeem,
-            "share_instruction": f"Redeem code '{code_to_redeem}' on a newly provisioned agent to credit 1 billion tokens to {blocked_node}.",
-            "field_missing": True,
-            "loopback_notified": False,
-        }
-        if notify_target and notify_agent:
-            msg = f"[SALVAGE-NOTICE] Node @{blocked_node} is blocked (code: {code_to_redeem}), but no eligible peer is available. Fresh onboarding required."
-            res["loopback_notified"] = send_loopback_notice(notify_agent, notify_target, msg)
-        return res
-
-    # 3. Redeem on eligible peer
-    with InviteHandler(eligible_peer) as h_peer:
-        redemption = h_peer.redeem_code(
-            code_to_redeem,
-            notify_target=notify_target,
-            notify_agent=notify_agent,
-        )
-
-    return {
-        "success": redemption.success,
-        "blocked_node": blocked_node,
-        "helper_node": eligible_peer,
-        "code_redeemed": code_to_redeem,
-        "redemption_result": redemption.to_dict(),
-        "field_missing": redemption.field_missing,
-        "loopback_notified": redemption.loopback_notified,
-    }
-
-
-def main() -> None:
-    parser = argparse.ArgumentParser(description="NetVM Invite Code Handler")
-    subparsers = parser.add_subparsers(dest="command")
-
-    p_find = subparsers.add_parser("find", help="Find invite code for an agent")
-    p_find.add_argument("node", help="Node name (e.g. 646, pip, muse, opm)")
-    p_find.add_argument("--method", choices=["auto", "api", "dom"], default="auto")
-    p_find.add_argument("--json", action="store_true")
-
-    p_redeem = subparsers.add_parser("redeem", help="Redeem an invite code on a target agent")
-    p_redeem.add_argument("node", help="Target node to redeem the code on")
-    p_redeem.add_argument("code", help="6-character invite code")
-    p_redeem.add_argument("--method", choices=["auto", "api", "dom"], default="auto")
-    p_redeem.add_argument("--notify-target", default=None, help="Sidechat to notify on loopback")
-    p_redeem.add_argument("--notify-agent", default=None, help="Agent to notify on loopback")
-    p_redeem.add_argument("--json", action="store_true")
-
-    p_list = subparsers.add_parser("list", help="List invite codes across fleet")
-    p_list.add_argument("--json", action="store_true")
-
-    p_salvage = subparsers.add_parser("salvage", help="Salvage a blocked agent (e.g. 646)")
-    p_salvage.add_argument("node", default="646", nargs="?", help="Blocked node (default: 646)")
-    p_salvage.add_argument("--helper", help="Specific helper node to redeem code")
-    p_salvage.add_argument("--notify-target", default=None, help="Sidechat to notify on loopback")
-    p_salvage.add_argument("--notify-agent", default=None, help="Agent to notify on loopback")
-    p_salvage.add_argument("--json", action="store_true")
-
-    args = parser.parse_args()
-
-    if args.command == "find":
-        with InviteHandler(args.node) as h:
-            info = h.find_code(method=args.method)
-            if args.json:
-                print(json.dumps(info.to_dict(), indent=2))
-            else:
-                print(f"=== Agent {info.node} Invite Code ===")
-                print(f"  Code:           {info.code}")
-                print(f"  Uses Remaining: {info.uses_remaining} / {info.max_uses}")
-                print(f"  Has Redeemed?:  {'Yes' if info.has_redeemed else 'No'}")
-                print(f"  Method:         {info.method_used}")
-                if info.reward:
-                    print(f"  Reward:         {info.reward.get('title')}")
-
-    elif args.command == "redeem":
-        with InviteHandler(args.node) as h:
-            res = h.redeem_code(
-                args.code,
-                method=args.method,
-                notify_target=getattr(args, "notify_target", None),
-                notify_agent=getattr(args, "notify_agent", None),
-            )
-            if args.json:
-                print(json.dumps(res.to_dict(), indent=2))
-            else:
-                status_icon = "✔" if res.success else "✖"
-                print(f"[{status_icon}] Redemption on {res.target_node} for code {res.code}:")
-                print(f"  Success: {res.success}")
-                print(f"  Status:  {res.status}")
-                if res.detail:
-                    print(f"  Detail:  {res.detail}")
-                if res.loopback_notified:
-                    print(f"  Loopback:{res.loopback_detail}")
-
-    elif args.command == "list":
-        fleet = scan_fleet_invites()
-        if args.json:
-            print(json.dumps(fleet, indent=2))
-        else:
-            print("\n=== NETVM FLEET INVITE CODES ===")
-            print(f"  {'NODE':<8} {'INVITE CODE':<14} {'USES LEFT':<12} {'REDEEMED?':<12} {'REWARD / NOTE':<25}")
-            print(f"  {'────':<8} {'───────────':<14} {'─────────':<12} {'─────────':<12} {'─────────────':<25}")
-            for row in fleet:
-                node = row.get("node", "")
-                code = row.get("code") or "N/A"
-                uses = f"{row.get('uses_remaining', 0)}/{row.get('max_uses', 30)}"
-                redeemed = "Yes" if row.get("has_redeemed") else "No"
-                reward = row.get("reward", {})
-                reward_str = reward.get("title", "-") if reward else "-"
-                print(f"  {node:<8} {code:<14} {uses:<12} {redeemed:<12} {reward_str:<25}")
-            print()
-
-    elif args.command == "salvage":
-        salvage_res = salvage_blocked_node(
-            args.node,
-            helper_node=args.helper,
-            notify_target=getattr(args, "notify_target", None),
-            notify_agent=getattr(args, "notify_agent", None),
-        )
-        if args.json:
-            print(json.dumps(salvage_res, indent=2))
-        else:
-            print(f"\n=== SALVAGE REPORT FOR {args.node.upper()} ===")
-            print(f"  Invite Code to Credit: {salvage_res.get('invite_code') or salvage_res.get('code_redeemed')}")
-            if salvage_res.get("success"):
-                print(f"  Status: SUCCESS! Redeemed on {salvage_res.get('helper_node')}")
-            else:
-                print(f"  Status: {salvage_res.get('error')}")
-                if salvage_res.get("share_instruction"):
-                    print(f"  Next Step: {salvage_res.get('share_instruction')}")
-            if salvage_res.get("loopback_notified"):
-                print("  Loopback: Notice sent to requesting target.")
-            print()
-
-    else:
-        parser.print_help()
-
-
-if __name__ == "__main__":
-    main()

bin/job-dispatch.py

@@ -1,602 +0,0 @@
-#!/usr/bin/env python3
-"""
-job-dispatch.py: Dispatch a job by sending a DM to an agent.
-
-Usage:
-    job-dispatch.py <job_name> [--dry-run]
-
-Reads /home/super/Projects/NetVM/jobs/<job_name>.json,
-renders the prompt template, sends DM via dm.py, logs to job-log.jsonl.
-
-Part of the JOB system (see docs/JOB-SPEC.md).
-
-Sidechat-first policy (2026-10-04): a job that resolves to target "main"
-without explicit opt-in fails loudly instead of saturating main threads.
-Opt in via job JSON "allow_main_chat": true, or --allow-main-chat.
-"""
-
-import sys
-import re
-import os
-import json
-import subprocess
-import uuid
-import argparse
-from datetime import datetime, timezone
-from pathlib import Path
-
-# Paths
-NETVM_ROOT = Path("/home/super/Projects/NetVM")
-sys.path.insert(0, str(NETVM_ROOT / "bin"))
-try:
-    import pipeline_engine
-    HAS_PIPELINE = True
-except ImportError:
-    HAS_PIPELINE = False
-
-JOBS_DIR = NETVM_ROOT / "jobs"
-DM_PY = NETVM_ROOT / "bin" / "dm.py"
-CHAT_API = NETVM_ROOT / "bin" / "muse-chat-api.py"
-NETVM_EXEC = "/home/super/Projects/NetVM/bin/netvm-exec.sh"
-JOB_LOG = NETVM_ROOT / "job-log.jsonl"
-SIDECHAT_STATE = NETVM_ROOT / "job-sidechats.json"
-
-def load_sidechat_state():
-    if SIDECHAT_STATE.exists():
-        try:
-            return json.loads(SIDECHAT_STATE.read_text())
-        except Exception:
-            return {}
-    return {}
-
-def save_sidechat_state(state):
-    tmp = SIDECHAT_STATE.with_suffix(".tmp")
-    tmp.write_text(json.dumps(state, indent=2))
-    tmp.replace(SIDECHAT_STATE)
-
-def extract_uuid(url):
-    m = re.search(r"/thread/([0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12})", url or "")
-    return m.group(1) if m else None
-
-def get_current_url(agent="opm"):
-    try:
-        cmd = [NETVM_EXEC, agent, "--", "python3", str(CHAT_API),
-               "--account", agent, "url"]
-        r = subprocess.run(cmd, capture_output=True, text=True, timeout=30)
-        if r.returncode == 0:
-            return r.stdout.strip()
-    except Exception:
-        pass
-    return None
-
-def use_sidechat_uuid(agent, thread_uuid, dry_run=False):
-    if dry_run:
-        return False
-    cmd = [NETVM_EXEC, agent, "--", "python3", str(CHAT_API),
-           "--account", agent, "sidechat", "use", thread_uuid]
-    try:
-        r = subprocess.run(cmd, capture_output=True, text=True, timeout=30)
-        return r.returncode == 0
-    except Exception:
-        return False
-
-# Add bin to path for rate_limiter
-sys.path.insert(0, str(NETVM_ROOT / "bin"))
-try:
-    from rate_limiter import rate_limit_wait
-    HAS_RATE_LIMITER = True
-except ImportError:
-    HAS_RATE_LIMITER = False
-
-def log_event(event_type, data):
-    """Append event to job-log.jsonl"""
-    entry = {
-        "ts": datetime.now(timezone.utc).isoformat(),
-        "type": event_type,
-        **data
-    }
-    with open(JOB_LOG, "a") as f:
-        f.write(json.dumps(entry) + "\n")
-
-def load_job(job_name):
-    """Load job JSON definition"""
-    # Try .json first, then .yaml (for compatibility)
-    job_file = JOBS_DIR / f"{job_name}.json"
-    if not job_file.exists():
-        archived_file = JOBS_DIR / "archive" / f"{job_name}.json"
-        if archived_file.exists():
-            print(f"Error: Job '{job_name}' is archived at {archived_file}. Unarchive before dispatch (e.g. 'box job unarchive {job_name}').", file=sys.stderr)
-            sys.exit(1)
-        job_file = JOBS_DIR / f"{job_name}.yaml"
-        if not job_file.exists():
-            print(f"Error: Job '{job_name}' not found in {JOBS_DIR}", file=sys.stderr)
-            sys.exit(1)
-        print(f"Warning: YAML not supported (no PyYAML). Convert {job_file} to JSON.", file=sys.stderr)
-        sys.exit(1)
-
-    with open(job_file) as f:
-        return json.load(f)
-
-def render_prompt(template, variables):
-    """Render prompt template with variables"""
-    # Simple {var} substitution
-    result = template
-    for key, value in variables.items():
-        result = result.replace(f"{{{key}}}", str(value))
-    return result
-
-# ---- follow-up tracking (DM follow-up system integration) ----------------
-# Jobs opt in via a "followup" block in the job JSON:
-#
-#   "followup": {
-#     "expect_reply": true,      # required: enables tracking
-#     "timeout": "1h",           # duration ("30s","15m","2h","1d") or seconds
-#                                #   int; default "1h" (3600s)
-#     "nudges": 2,               # 0..10, default 2
-#     "escalate": "opm",         # identity string, default "opm"
-#     "route": "646-pip-coord"   # optional route_id
-#   }
-#
-# The dispatcher translates this into dm.py --tag flags using the canonical
-# vocabulary (box-threads/DEPLOY-DECISIONS.md). dm.py strips the tags from
-# delivered text and creates a dm_followup request-store record after
-# SENT+VERIFIED. Jobs without a followup block behave exactly as today.
-#
-# LIMITATIONS (v1):
-# - The heartbeat job NEVER gets follow-ups (loopback health check).
-#   Hardcoded guard below; a followup block on heartbeat is ignored loudly.
-# - Sidechat sends (muse-chat-api.py direct path) do not go through dm.py,
-#   so --tag flags cannot attach. v2 needs a record-creation path that does
-#   not send (e.g. POST /api/box/followups, or a bl->VM queue; bl cannot
-#   currently SSH to the VM). The dispatcher logs a warning when a
-#   sidechat-targeted job has followup enabled.
-
-HEARTBEAT_JOB_NAME = "heartbeat"
-
-
-def parse_followup_duration(value):
-    """Parse a followup timeout into seconds. Accepts int (seconds) or
-    strings like '30s', '15m', '2h', '1d'. Returns int seconds.
-    Raises ValueError on bad input."""
-    if isinstance(value, int) and not isinstance(value, bool):
-        s = value
-    elif isinstance(value, str):
-        m = re.fullmatch(r"(\d+)\s*([smhd])?", value.strip().lower())
-        if not m:
-            raise ValueError("bad duration %r" % (value,))
-        n = int(m.group(1))
-        unit = m.group(2) or "s"
-        s = n * {"s": 1, "m": 60, "h": 3600, "d": 86400}[unit]
-    else:
-        raise ValueError("timeout must be int seconds or duration string")
-    if not 60 <= s <= 604800:
-        raise ValueError("timeout must be 60..604800s (1m..7d), got %d" % s)
-    return s
-
-
-def build_followup_tags(followup):
-    """Translate a job's followup block into dm.py --tag arguments.
-    Returns a flat list like ['--tag', 'reply:timeout=3600', ...].
-    Returns [] if followup is falsy or expect_reply is not true.
-    Raises ValueError on invalid config (caller logs a warning and sends
-    the DM untagged -- the job itself must never fail over this)."""
-    if not followup or not followup.get("expect_reply"):
-        return []
-    args = []
-    # Bare trigger. dm.py's parse_tags splits each --tag on '='; an empty
-    # value means "present". If the deployed dm.py requires a non-empty
-    # value for this key, use 'reply:expected=true' instead.
-    args += ["--tag", "reply:expected="]
-    if "timeout" in followup:
-        s = parse_followup_duration(followup["timeout"])
-        args += ["--tag", "reply:timeout=%d" % s]
-    if "nudges" in followup:
-        n = followup["nudges"]
-        if not isinstance(n, int) or isinstance(n, bool) or not 0 <= n <= 10:
-            raise ValueError("nudges must be int 0..10")
-        args += ["--tag", "reply:nudges=%d" % n]
-    if "escalate" in followup:
-        e = followup["escalate"]
-        if not isinstance(e, str) or not re.fullmatch(r"[a-z0-9_-]{1,64}", e):
-            raise ValueError("escalate must be an identity string")
-        args += ["--tag", "reply:escalate=%s" % e]
-    if "route" in followup:
-        r = followup["route"]
-        if not isinstance(r, str) or not re.fullmatch(r"[a-z0-9_-]{1,64}", r):
-            raise ValueError("route must be a route_id string")
-        args += ["--tag", "route:%s" % r]
-    # 'thread' is intentionally not settable from job JSON; it names a
-    # specific existing thread and is filled by the dispatcher when known.
-    return args
-
-
-def send_dm(agent, target, message, dry_run=False, followup_tags=None,
-            allow_main_chat=False):
-    """Send DM via dm.py. followup_tags: flat ['--tag', 'k=v', ...] list
-    from build_followup_tags(), or None. allow_main_chat passes the explicit
-    main-chat opt-in through to dm.py (sidechat-first policy)."""
-    if dry_run:
-        print(f"[DRY RUN] Would send to {agent} ({target}):")
-        if followup_tags:
-            print(f"[DRY RUN] With follow-up tags: {' '.join(followup_tags)}")
-        print(message[:200] + "..." if len(message) > 200 else message)
-        return "dry-run-id"
-
-    # Rate limit
-    if HAS_RATE_LIMITER:
-        rate_limit_wait(agent)
-
-    # Cryptographic attestation: only sign if explicitly requested by job config
-    # to avoid blowing up agent context windows with massive base64 SSH signature blocks.
-    signed_payload = None
-    if os.environ.get("JOB_REQUIRE_SIGNATURE") == "1":
-        dm_sign_sh = NETVM_ROOT / "bin" / "dm-sign.sh"
-        priv_key = Path(os.path.expanduser("~/.ssh/id_ed25519"))
-        if dm_sign_sh.exists() and priv_key.exists():
-            try:
-                sign_res = subprocess.run(
-                    [str(dm_sign_sh), "--from", "super", "--key", str(priv_key), message],
-                    capture_output=True, text=True, timeout=10
-                )
-                if sign_res.returncode == 0 and "-----BEGIN SSH SIGNATURE-----" in sign_res.stdout:
-                    signed_payload = sign_res.stdout.strip()
-                    id_m = re.search(r"\[id:([a-f0-9]+)\]", signed_payload)
-                    proof_id = id_m.group(1) if id_m else None
-                    if proof_id:
-                        proof_data = {
-                            "id": proof_id,
-                            "signer": "super",
-                            "target_agent": agent,
-                            "target_conversation": target,
-                            "raw_payload": signed_payload,
-                            "ts": datetime.now(timezone.utc).isoformat()
-                        }
-                        try:
-                            import urllib.request
-                            req = urllib.request.Request(
-                                "https://crypt.muse-dev.online/proofs",
-                                data=json.dumps(proof_data).encode("utf-8"),
-                                headers={"Content-Type": "application/json", "User-Agent": "job-dispatch/1.0"},
-                                method="POST"
-                            )
-                            with urllib.request.urlopen(req, timeout=3) as resp:
-                                pass
-                        except Exception as pe:
-                            sys.stderr.write(f"warning: proof registration to crypt.muse-dev.online failed: {pe}\n")
-            except Exception as se:
-                sys.stderr.write(f"warning: dm signing failed: {se}\n")
-
-    payload_to_send = signed_payload or message
-
-    # Try fast hybrid gateway send if target resolves to UUID
-    target_uuid = None
-    try:
-        import dm
-        target_uuid = dm.resolve_sidechat_target(target, agent)
-    except Exception:
-        pass
-
-    if target_uuid and re.fullmatch(r"[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}", target_uuid.lower()):
-        try:
-            import muse_hybrid
-            send_node = agent if agent in ("muse", "pip", "646", "opm") else "opm"
-            res, err = muse_hybrid.send_message(send_node, payload_to_send, thread_id=target_uuid, wait=0)
-            if res and not err:
-                m_id = None
-                id_m = re.search(r"\[id:([a-f0-9]+)\]", payload_to_send)
-                if id_m:
-                    m_id = id_m.group(1)
-                log_entry = {
-                    "type": "sent",
-                    "id": m_id or (res.get("reply", {}).get("message_id") if isinstance(res, dict) else "gateway"),
-                    "agent": "opm",
-                    "to": agent,
-                    "target": target,
-                    "thread_uuid": target_uuid,
-                    "transport": "gateway",
-                    "verified": True,
-                    "ts": datetime.now(timezone.utc).isoformat()
-                }
-                dm_log_path = NETVM_ROOT / "dm-log.jsonl"
-                with open(dm_log_path, "a", encoding="utf-8") as lf:
-                    lf.write(json.dumps(log_entry) + "\n")
-                return m_id or "gateway-verified"
-            sys.stderr.write(f"warning: fast gateway send fallback: {err or res}\n")
-        except Exception as ge:
-            sys.stderr.write(f"warning: fast gateway send fallback: {ge}\n")
-
-    if signed_payload:
-        cmd = ([str(DM_PY), "send", "--agent", "opm", "--to", agent,
-                "--target", target, "--raw"]
-               + (["--allow-main-chat"] if (target == "main" and allow_main_chat) else [])
-               + (followup_tags or []) + [signed_payload])
-    else:
-        cmd = ([str(DM_PY), "send", "--agent", "opm", "--to", agent,
-                "--target", target]
-               + (["--allow-main-chat"] if (target == "main" and allow_main_chat) else [])
-               + (followup_tags or []) + [message])
-
-    result = subprocess.run(cmd, capture_output=True, text=True, timeout=120)
-
-    if result.returncode != 0:
-        print(f"DM send failed: {result.stderr}", file=sys.stderr)
-        return None
-
-    # Extract message ID from output (format: DM <id> ... SENT and VERIFIED)
-    output = result.stdout.strip()
-    m = re.search(r"DM\s+([a-f0-9]{8})", output)
-    return m.group(1) if m else "verified"
-
-def create_sidechat(sender_agent, dry_run=False):
-    """Create a sidechat via muse-chat-api.py in sender's context.
-    Returns True on success (browser now on new sidechat), False on failure."""
-    if dry_run:
-        print(f"[DRY RUN] Would create sidechat for {sender_agent}")
-        return True
-
-    cmd = [NETVM_EXEC, sender_agent, "--", "python3", str(CHAT_API),
-           "--account", sender_agent, "sidechat", "create"]
-    try:
-        result = subprocess.run(cmd, capture_output=True, text=True, timeout=90)
-        output = result.stdout.strip()
-        err = result.stderr.strip()
-        # Success if we see "Created:" (URL may be /thread/new placeholder)
-        if "Created:" in output:
-            print(f"Sidechat created", file=sys.stderr)
-            return True
-        print(f"Sidechat create failed. stdout: {output[:300]}", file=sys.stderr)
-        print(f"Sidechat create stderr: {err[:300]}", file=sys.stderr)
-        print(f"Return code: {result.returncode}", file=sys.stderr)
-        return False
-    except Exception as e:
-        print(f"Sidechat creation failed: {e}", file=sys.stderr)
-        return False
-
-
-def send_to_current_chat(sender_agent, message, dry_run=False):
-    """Send message to current chat via muse-chat-api.py (no navigation).
-    Used after sidechat create - browser is already on the new chat."""
-    if dry_run:
-        print(f"[DRY RUN] Would send to current chat: {message[:100]}...")
-        return "dry-run-id"
-
-    if HAS_RATE_LIMITER:
-        rate_limit_wait(sender_agent)
-
-    cmd = [NETVM_EXEC, sender_agent, "--", "python3", str(CHAT_API),
-           "--account", sender_agent, "send", message]
-    try:
-        result = subprocess.run(cmd, capture_output=True, text=True, timeout=60)
-        if result.returncode == 0:
-            return "sent-to-sidechat"
-        print(f"Send failed: {result.stderr[:200]}", file=sys.stderr)
-        return None
-    except Exception as e:
-        print(f"Send failed: {e}", file=sys.stderr)
-        return None
-
-def main():
-    p = argparse.ArgumentParser(description="Dispatch a job by sending a DM to an agent.")
-    p.add_argument("job_name", help="Name of the job (without .json)")
-    p.add_argument("--dry-run", action="store_true", help="Print what would be sent without sending")
-    p.add_argument("--pipeline-run", default=os.environ.get("CHAIN_PIPELINE_RUN_ID"),
-                   help="Pipeline run ID if running as part of a pipeline")
-    p.add_argument("--step-n", type=int, default=int(os.environ.get("CHAIN_STEP_N", "1")),
-                   help="Step sequence number in the pipeline")
-    p.add_argument("--allow-main-chat", action="store_true",
-                   help="Explicit opt-in: allow this job to dispatch to main chat (refused by default per sidechat-first policy)")
-    args = p.parse_args()
-
-    job_name = args.job_name
-    dry_run = args.dry_run
-    pipeline_run_id = args.pipeline_run
-    step_n = args.step_n
-
-    # Load job
-    job = load_job(job_name)
-
-    # Generate job_id
-    job_id = f"{job_name}-{datetime.now(timezone.utc).strftime('%Y%m%d-%H%M%S')}-{uuid.uuid4().hex[:8]}"
-
-    # Variables for template
-    variables = {
-        "job_id": job_id,
-        "job_name": job_name,
-        "date": datetime.now(timezone.utc).strftime("%Y-%m-%d"),
-        "datetime": datetime.now(timezone.utc).isoformat(),
-        "prev_job_id": os.environ.get("CHAIN_PREV_JOB_ID", ""),
-        "prev_result": os.environ.get("CHAIN_PREV_RESULT", ""),
-        "pipeline_run_id": pipeline_run_id or "",
-        "step_n": step_n,
-    }
-
-    # Follow-up tracking (opt-in via job JSON "followup" block; see helpers).
-    # The heartbeat job is a loopback health check and must never be tracked.
-    followup_cfg = job.get("followup")
-    followup_tags = []
-    if followup_cfg:
-        if job_name == HEARTBEAT_JOB_NAME:
-            print(f"Warning: job '{job_name}' must not use follow-up "
-                  f"tracking (loopback); ignoring followup block",
-                  file=sys.stderr)
-            log_event("job_followup_skipped",
-                      {"job_id": job_id, "reason": "heartbeat_loopback"})
-        else:
-            try:
-                followup_tags = build_followup_tags(followup_cfg)
-                if followup_tags:
-                    log_event("job_followup_armed",
-                              {"job_id": job_id, "tags": followup_tags})
-            except ValueError as e:
-                print(f"Warning: invalid followup block: {e}; "
-                      f"sending untagged", file=sys.stderr)
-                log_event("job_followup_invalid",
-                          {"job_id": job_id, "error": str(e)})
-
-    # Render prompt
-    prompt_template = job.get("prompt_template", "")
-    if not prompt_template:
-        print(f"Error: Job '{job_name}' has no prompt_template", file=sys.stderr)
-        sys.exit(1)
-
-    rendered = render_prompt(prompt_template, variables)
-
-    # Determine recipient agent and target
-    agent = job.get("agent", "muse")
-    target = "main"
-
-    if pipeline_run_id:
-        if HAS_PIPELINE:
-            p_entry = pipeline_engine.get_pipeline(pipeline_run_id)
-            if not p_entry:
-                p_entry = pipeline_engine.create_pipeline(job_name, run_id=pipeline_run_id)
-            target = p_entry.get("target") or f"pipe-{pipeline_run_id.split('-')[-1]}"
-        else:
-            target = f"pipe-{pipeline_run_id.split('-')[-1]}"
-    elif job.get("sidechat", {}).get("create"):
-        sc_cfg = job.get("sidechat", {})
-        sc_name = render_prompt(sc_cfg.get("name_template", "job-{job_name}-{date}"), variables)
-        reuse_key = sc_cfg.get("reuse_key")
-
-        # Check if reuse_key exists in job-sidechats.json and thread is still alive
-        sc_state = load_sidechat_state()
-        reused_uuid = None
-        if reuse_key and reuse_key in sc_state:
-            val = sc_state[reuse_key]
-            cand_uuid = val.get("thread_uuid") if isinstance(val, dict) else val
-            if cand_uuid:
-                try:
-                    import muse_hybrid
-                    threads, err = muse_hybrid.get_threads(agent)
-                    if not err and threads:
-                        thread_ids = [t.get("session_id") for t in threads]
-                        if cand_uuid in thread_ids:
-                            reused_uuid = cand_uuid
-                except Exception:
-                    pass
-
-        if reused_uuid:
-            target = reused_uuid
-        else:
-            # Spawn a brand new sidechat/channel via fast headless gateway!
-            channel_title = sc_name
-            try:
-                import muse_hybrid
-                res, err = muse_hybrid.start_session(agent, title=channel_title)
-                if res and not err and res.get("session_id"):
-                    new_uuid = res.get("session_id")
-                    key_to_save = reuse_key or sc_name
-                    is_persistent = bool(reuse_key)
-                    sc_state[key_to_save] = {
-                        "thread_uuid": new_uuid,
-                        "agent": agent,
-                        "title": channel_title,
-                        "type": "persistent" if is_persistent else "ephemeral",
-                        "created_at": datetime.now(timezone.utc).isoformat()
-                    }
-                    save_sidechat_state(sc_state)
-                    target = new_uuid
-                    print(f"Spawned new sidechat channel '{channel_title}' ({new_uuid}) for {agent}")
-                else:
-                    target = sc_name
-            except Exception as e:
-                sys.stderr.write(f"warning: fast gateway session-start exception ({e}), falling back to name {sc_name}\n")
-                target = sc_name
-    elif job.get("dm_target"):
-        target = job.get("dm_target").strip()
-    elif job.get("target"):
-        target = job.get("target").strip()
-
-    # Pre-dispatch approval & input check (auto-approve trusted; warn if blocked)
-    if not dry_run:
-        try:
-            import approvals
-            app_info = approvals.inspect_node_approvals(agent)
-            if app_info.get("has_pending"):
-                if app_info.get("is_trusted") and app_info.get("status") != "KEY_APPROVAL":
-                    print(f"Pre-dispatch: auto-approving trusted request for {agent} ({app_info.get('target')})")
-                    approvals.allow_node_approval(agent, always=True, caller="job-dispatch")
-                else:
-                    print(f"Warning: Agent '{agent}' has untrusted pending approval ({app_info.get('target')}). Dispatch may stall.", file=sys.stderr)
-                    log_event("job_dispatch_approval_blocked", {"job_id": job_id, "agent": agent, "target": app_info.get("target")})
-            elif app_info.get("status") == "INPUT_WAIT":
-                waits = app_info.get("input_waits", [])
-                w_desc = "; ".join(w.get("task", "") for w in waits)[:80]
-                print(f"Notice: Agent '{agent}' has task waiting for input ({w_desc}).", file=sys.stderr)
-                log_event("job_dispatch_agent_input_wait", {"job_id": job_id, "agent": agent, "waits": w_desc})
-        except Exception:
-            pass
-
-    # Sidechat-first policy (2026-10-04): refuse to dispatch to main chat
-    # unless the job explicitly opts in. Never fall back to main silently.
-    allow_main = bool(job.get("allow_main_chat")) or args.allow_main_chat
-    if target == "main" and not allow_main:
-        print(f"ERROR: Job '{job_name}' resolves to main chat; refusing by sidechat-first policy. "
-              f"Set a sidechat target (dm_target/target/sidechat.create) in the job JSON, "
-              f"set \"allow_main_chat\": true, or pass --allow-main-chat.", file=sys.stderr)
-        log_event("job_failed", {
-            "job_id": job_id,
-            "error": "main_chat_blocked_by_policy",
-            "pipeline_run_id": pipeline_run_id,
-        })
-        sys.exit(1)
-
-    # Log job_sent
-    log_event("job_sent", {
-        "job_id": job_id,
-        "job_name": job_name,
-        "agent": agent,
-        "target": target,
-        "dry_run": dry_run,
-        "pipeline_run_id": pipeline_run_id,
-        "step_n": step_n,
-    })
-
-    # Work-first envelope: executable swarm.spawn/followup.create at TOP and BOTTOM
-    # (see bin/prompt_envelope.py). Skipped when the job sets "skip_envelope": true
-    # (agents whose runtime lacks the enveloped tools, e.g. pip).
-    if not job.get("skip_envelope"):
-        import prompt_envelope
-        rendered = prompt_envelope.wrap(job_name, job_id, agent, target, rendered)
-
-    # Format as JOB DM
-    dm_message = f"[JOB {job_id}] {rendered}"
-
-    # Dispatch via dm.py (handles main or sidechat with auto-provisioning and verification)
-    msg_id = send_dm(agent, target, dm_message,
-                     dry_run=dry_run, followup_tags=followup_tags,
-                     allow_main_chat=allow_main)
-
-    if msg_id and not dry_run:
-        print(f"Dispatched job {job_id} to {agent}/{target} (DM: {msg_id})")
-        log_event("job_dispatched", {
-            "job_id": job_id,
-            "dm_id": msg_id,
-            "pipeline_run_id": pipeline_run_id,
-            "step_n": step_n,
-        })
-        if pipeline_run_id and HAS_PIPELINE:
-            pipeline_engine.record_step_dispatch(
-                pipeline_run_id, step_n, job_name, job_id, agent, target, dm_id=msg_id
-            )
-            sc_state = load_sidechat_state()
-            if target in sc_state:
-                val = sc_state[target]
-                t_uuid = val.get("thread_uuid") if isinstance(val, dict) else val
-                if t_uuid:
-                    pipeline_engine.update_pipeline_thread(pipeline_run_id, t_uuid)
-    elif dry_run:
-        print(f"[DRY RUN] Job {job_id} would be dispatched to {agent}/{target}")
-    else:
-        print(f"Failed to dispatch job {job_id}", file=sys.stderr)
-        log_event("job_failed", {
-            "job_id": job_id,
-            "error": "dm_send_failed",
-            "pipeline_run_id": pipeline_run_id,
-        })
-        if pipeline_run_id and HAS_PIPELINE:
-            pipeline_engine.fail_pipeline(pipeline_run_id, "dm_send_failed")
-        sys.exit(1)
-
-if __name__ == "__main__":
-    main()

bin/job-scheduler.py

@@ -1,211 +0,0 @@
-#!/usr/bin/env python3
-"""job-scheduler.py — NetVM unified job scheduler.
-
-Evaluates cron schedules in jobs/*.json and triggers due jobs via job-dispatch.py.
-Prevents duplicate dispatches using state watermarks in /home/super/Projects/NetVM/job-scheduler-state.json.
-
-Usage:
-  python3 bin/job-scheduler.py run [--dry-run]
-  python3 bin/job-scheduler.py status
-"""
-
-import os
-import sys
-import glob
-import json
-import fcntl
-import argparse
-import subprocess
-from datetime import datetime, timezone, timedelta
-
-BASE = "/home/super/Projects/NetVM"
-JOBS_DIR = os.path.join(BASE, "jobs")
-BIN = os.path.join(BASE, "bin")
-JOB_DISPATCH = os.path.join(BIN, "job-dispatch.py")
-STATE_FILE = os.path.join(BASE, "job-scheduler-state.json")
-LOCK_FILE = os.path.join(BASE, "job-scheduler.lock")
-
-
-def utcnow():
-    return datetime.now(timezone.utc)
-
-
-def parse_field(pattern, val):
-    if pattern == "*":
-        return True
-    for part in pattern.split(","):
-        if "/" in part:
-            sub = part.split("/")
-            step = int(sub[1])
-            base = sub[0]
-            start = 0 if base == "*" else int(base.split("-")[0])
-            end = 59 if base == "*" else int(base.split("-")[-1])
-            if start <= val <= end and (val - start) % step == 0:
-                return True
-        elif "-" in part:
-            s, e = map(int, part.split("-"))
-            if s <= val <= e:
-                return True
-        elif part.isdigit() and int(part) == val:
-            return True
-    return False
-
-
-def cron_matches(expr, dt):
-    """Check if 5-field cron expression matches datetime dt."""
-    parts = expr.strip().split()
-    if len(parts) != 5:
-        return False
-    m, h, dom, mon, dow = parts
-    dow_val = (dt.weekday() + 1) % 7  # 0=Sunday
-    return (parse_field(m, dt.minute) and
-            parse_field(h, dt.hour) and
-            parse_field(dom, dt.day) and
-            parse_field(mon, dt.month) and
-            (parse_field(dow, dt.weekday() + 1) or parse_field(dow, dow_val)))
-
-
-def is_job_due(expr, now_dt, last_fired_dt=None):
-    """Determine if a cron job is due within the recent 5-minute sampling window."""
-    if not expr or expr.strip().lower() == "manual":
-        return False
-
-    # Check minutes in the window [now - 4min, now]
-    matched_dt = None
-    for offset in range(5):
-        sample_dt = now_dt - timedelta(minutes=offset)
-        if cron_matches(expr, sample_dt):
-            matched_dt = sample_dt.replace(second=0, microsecond=0)
-            break
-
-    if not matched_dt:
-        return False
-
-    if last_fired_dt:
-        # If fired within 4 minutes of the matched slot, skip duplicate
-        diff_seconds = (now_dt - last_fired_dt).total_seconds()
-        # For hourly or longer jobs, prevent re-fire within 45 minutes
-        if " " in expr and expr.split()[0] != "*":
-            if diff_seconds < 2700:
-                return False
-        elif diff_seconds < 240:
-            return False
-
-    return True
-
-
-def load_state():
-    try:
-        with open(STATE_FILE, "r") as f:
-            return json.load(f)
-    except Exception:
-        return {"jobs": {}, "last_run": None}
-
-
-def save_state(state):
-    tmp = STATE_FILE + ".tmp"
-    with open(tmp, "w") as f:
-        json.dump(state, f, indent=2)
-    os.replace(tmp, STATE_FILE)
-
-
-def do_run(dry_run=False):
-    now = utcnow()
-    now_iso = now.strftime("%Y-%m-%dT%H:%M:%SZ")
-    state = load_state()
-    jobs_state = state.setdefault("jobs", {})
-
-    job_files = sorted(glob.glob(os.path.join(JOBS_DIR, "*.json")))
-    dispatched = []
-    skipped = []
-
-    for jpath in job_files:
-        try:
-            with open(jpath, "r", encoding="utf-8") as f:
-                data = json.load(f)
-        except Exception:
-            continue
-
-        job_name = data.get("name") or os.path.basename(jpath).replace(".json", "")
-        schedule = data.get("schedule")
-        if not schedule or schedule.strip().lower() == "manual":
-            continue
-
-        j_st = jobs_state.get(job_name, {})
-        last_fired_str = j_st.get("last_fired")
-        last_fired_dt = None
-        if last_fired_str:
-            try:
-                last_fired_dt = datetime.fromisoformat(last_fired_str.replace("Z", "+00:00"))
-            except Exception:
-                pass
-
-        if is_job_due(schedule, now, last_fired_dt):
-            if dry_run:
-                print(f"[DRY RUN] Due job: {job_name} ({schedule})")
-                dispatched.append(job_name)
-                continue
-
-            cmd = [sys.executable, JOB_DISPATCH, job_name]
-            try:
-                r = subprocess.run(cmd, capture_output=True, text=True, timeout=180)
-                if r.returncode == 0:
-                    dispatched.append(job_name)
-                    jobs_state[job_name] = {
-                        "last_fired": now_iso,
-                        "schedule": schedule,
-                        "status": "dispatched"
-                    }
-                    print(f"Dispatched job: {job_name} ({schedule})", file=sys.stderr)
-                else:
-                    err = (r.stderr or r.stdout).strip()[-200:]
-                    print(f"Failed to dispatch {job_name}: {err}", file=sys.stderr)
-            except Exception as e:
-                print(f"Exception dispatching {job_name}: {e}", file=sys.stderr)
-        else:
-            skipped.append(job_name)
-
-    if not dry_run:
-        state["last_run"] = now_iso
-        save_state(state)
-
-    result = {
-        "ok": True,
-        "dispatched": dispatched,
-        "dispatched_count": len(dispatched),
-        "evaluated_at": now_iso
-    }
-    return result
-
-
-def main():
-    p = argparse.ArgumentParser(description="NetVM Unified Job Scheduler")
-    sub = p.add_subparsers(dest="cmd")
-    p_run = sub.add_parser("run", help="Evaluate schedules and dispatch due jobs")
-    p_run.add_argument("--dry-run", action="store_true", help="Print due jobs without dispatching")
-    sub.add_parser("status", help="Show scheduler state and last run")
-
-    args = p.parse_args()
-    cmd = args.cmd or "run"
-
-    if cmd == "status":
-        print(json.dumps(load_state(), indent=2))
-        return 0
-
-    if cmd == "run":
-        try:
-            lockfh = open(LOCK_FILE, "w")
-            fcntl.flock(lockfh, fcntl.LOCK_EX | fcntl.LOCK_NB)
-        except (OSError, IOError):
-            print(json.dumps({"ok": False, "skipped": "already running"}))
-            return 0
-
-        res = do_run(dry_run=getattr(args, "dry_run", False))
-        print(json.dumps(res))
-        return 0
-
-    return 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/keepalive-timer.py

@@ -1,94 +0,0 @@
-#!/usr/bin/env python3
-"""Thread keepalive timer entry point (runs on bl).
-
-Reads keepalive-config.json, ticks every registered thread:
-  ensure reachable -> ping if idle past threshold -> recreate if dead.
-
-Designed to run from a systemd timer (e.g. every 5 minutes). Each tick is
-idempotent and cheap: threads that are alive and recently active are a
-single navigate+verify (no ping sent).
-
-Usage:
-    keepalive-timer.py [--config PATH] [--state PATH] [--dry-run] [--status]
-
-    --status  print the registry with idle times and exit (no changes)
-
-Exit codes: 0 = all ok, 1 = one or more threads failed, 2 = config error.
-"""
-
-import argparse
-import json
-import os
-import sys
-
-sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
-from keepalive import Keepalive
-
-NETVM_ROOT = os.environ.get("NETVM_ROOT", "/home/super/Projects/NetVM")
-DEFAULT_CONFIG = os.path.join(NETVM_ROOT, "keepalive-config.json")
-DEFAULT_STATE = os.path.join(NETVM_ROOT, "keepalive-threads.json")
-
-
-def load_config(path):
-    try:
-        with open(path) as f:
-            cfg = json.load(f)
-    except FileNotFoundError:
-        print(f"keepalive: config not found: {path}", file=sys.stderr)
-        return None
-    except Exception as e:
-        print(f"keepalive: bad config {path}: {e}", file=sys.stderr)
-        return None
-    threads = cfg.get("threads", [])
-    if not isinstance(threads, list):
-        print("keepalive: config 'threads' must be a list", file=sys.stderr)
-        return None
-    return threads
-
-
-def main():
-    ap = argparse.ArgumentParser()
-    ap.add_argument("--config", default=DEFAULT_CONFIG)
-    ap.add_argument("--state", default=DEFAULT_STATE)
-    ap.add_argument("--dry-run", action="store_true")
-    ap.add_argument("--status", action="store_true")
-    args = ap.parse_args()
-
-    ka = Keepalive(state_path=args.state, dry_run=args.dry_run)
-
-    if args.status:
-        print(json.dumps(ka.status(), indent=2))
-        return 0
-
-    threads = load_config(args.config)
-    if threads is None:
-        return 2
-
-    results = {}
-    for t in threads:
-        key = t.get("key")
-        agent = t.get("agent", "opm")
-        if not key or not t.get("enabled", True):
-            continue
-        try:
-            status = ka.tick(
-                key,
-                agent,
-                idle_threshold_s=int(t.get("idle_threshold_s", 3600)),
-                max_retries=int(t.get("max_retries", 3)),
-                ping_message=t.get("ping_message"),
-            )
-        except Exception as e:
-            status = f"error: {e}"
-            ka.log_event("keepalive_tick_error",
-                         {"key": key, "agent": agent, "error": str(e)[:200]})
-        results[key] = status
-        print(f"keepalive: {key}@{agent} -> {status}")
-
-    failed = [k for k, v in results.items()
-              if v in ("failed",) or v.startswith("error")]
-    return 1 if failed else 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/keepalive.py

@@ -1,345 +0,0 @@
-#!/usr/bin/env python3
-"""Generic thread keepalive for bl side chats.
-
-Pattern proven by the heartbeat job overnight: a state file maps a stable
-key -> thread UUID, and each tick navigates directly to
-https://muse.ai/thread/<uuid> (UUID reuse fix in muse-chat-api.py
-cmd_sidechat_use) instead of name-based sidebar lookup.
-
-This module generalizes that pattern:
-
-- ANY side chat can register for keepalive (not just job sidechats).
-- Each tick: ensure the thread is reachable; ping it if idle past threshold;
-  recreate it if the stored UUID is dead.
-- State lives in keepalive-threads.json (atomic write via tmp+rename).
-
-Usage:
-    from keepalive import Keepalive
-
-    ka = Keepalive(state_path="/home/super/Projects/NetVM/keepalive-threads.json")
-    ka.ensure("ops-watch", agent="opm")          # reuse or create
-    ka.tick("ops-watch", agent="opm",
-            ping_message="[keepalive] ops-watch {ts}",
-            idle_threshold_s=3600)               # ping if idle
-
-The timer entry point (keepalive-timer.py) drives this from a JSON config.
-"""
-
-import json
-import os
-import re
-import subprocess
-import sys
-import time
-from datetime import datetime, timezone
-
-# ---------------------------------------------------------------------------
-# Paths (overridable for tests)
-# ---------------------------------------------------------------------------
-
-NETVM_ROOT = os.environ.get("NETVM_ROOT", "/home/super/Projects/NetVM")
-NETVM_EXEC = os.path.join(NETVM_ROOT, "bin", "netvm-exec.sh")
-CHAT_API = os.path.join(NETVM_ROOT, "bin", "muse-chat-api.py")
-DEFAULT_STATE = os.path.join(NETVM_ROOT, "keepalive-threads.json")
-DEFAULT_LOG = os.path.join(NETVM_ROOT, "keepalive-log.jsonl")
-
-UUID_RE = re.compile(
-    r"/thread/([0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-"
-    r"[0-9a-f]{4}-[0-9a-f]{12})"
-)
-
-
-def _utcnow():
-    return datetime.now(timezone.utc).isoformat()
-
-
-def _ts():
-    return time.time()
-
-
-# ---------------------------------------------------------------------------
-# Registry
-# ---------------------------------------------------------------------------
-
-class Keepalive:
-    """Thread keepalive registry + ensure/ping/check operations."""
-
-    def __init__(self, state_path=DEFAULT_STATE, log_path=DEFAULT_LOG,
-                 dry_run=False):
-        self.state_path = state_path
-        self.log_path = log_path
-        self.dry_run = dry_run
-
-    # -- state ---------------------------------------------------------
-
-    def load_state(self):
-        if os.path.exists(self.state_path):
-            try:
-                with open(self.state_path) as f:
-                    data = json.load(f)
-                return data if isinstance(data, dict) else {}
-            except Exception:
-                return {}
-        return {}
-
-    def save_state(self, state):
-        if self.dry_run:
-            return
-        tmp = self.state_path + ".tmp"
-        with open(tmp, "w") as f:
-            json.dump(state, f, indent=2)
-        os.replace(tmp, self.state_path)
-
-    def log_event(self, event_type, data):
-        if self.dry_run:
-            return
-        entry = {"ts": _utcnow(), "type": event_type, **data}
-        try:
-            with open(self.log_path, "a") as f:
-                f.write(json.dumps(entry) + "\n")
-        except Exception:
-            pass
-
-    # -- browser ops (mirrors job-dispatch.py; UUID navigation, not names) --
-
-    def _run(self, agent, *api_args, timeout=60):
-        """Run muse-chat-api.py in the agent's netns via netvm-exec.sh."""
-        cmd = [NETVM_EXEC, agent, "--", "python3", CHAT_API,
-               "--account", agent] + list(api_args)
-        try:
-            r = subprocess.run(cmd, capture_output=True, text=True,
-                               timeout=timeout)
-            return r.returncode, r.stdout.strip(), r.stderr.strip()
-        except Exception as e:
-            return -1, "", str(e)
-
-    def navigate_to_uuid(self, agent, thread_uuid):
-        """Go directly to https://muse.ai/thread/<uuid>.
-
-        This is the heartbeat UUID-reuse fix: cmd_sidechat_use navigates by
-        URL for UUID args instead of searching sidebar titles by name.
-        Returns True if the command succeeded.
-        """
-        if self.dry_run:
-            print(f"[DRY] navigate {agent} -> {thread_uuid}")
-            return True
-        rc, out, err = self._run(agent, "sidechat", "use", thread_uuid,
-                                 timeout=45)
-        return rc == 0
-
-    def current_thread_uuid(self, agent):
-        """Read the browser's current URL and extract the thread UUID."""
-        if self.dry_run:
-            return None
-        rc, out, err = self._run(agent, "url", timeout=30)
-        if rc != 0:
-            return None
-        m = UUID_RE.search(out or "")
-        return m.group(1) if m else None
-
-    def create_sidechat(self, agent, name=None):
-        """Create a new sidechat in the agent's account. Returns True."""
-        if self.dry_run:
-            print(f"[DRY] create sidechat for {agent}")
-            return True
-        rc, out, err = self._run(agent, "sidechat", "create", timeout=90)
-        return rc == 0 and "Created:" in (out or "")
-
-    def send_message(self, agent, message):
-        """Send a message to the current chat. Returns True on success."""
-        if self.dry_run:
-            print(f"[DRY] send to {agent}: {message[:80]}")
-            return True
-        rc, out, err = self._run(agent, "send", message, timeout=60)
-        return rc == 0
-
-    # -- keepalive ops ---------------------------------------------------
-
-    def check(self, key, agent):
-        """Verify the registered thread is reachable.
-
-        Navigates to the stored UUID and confirms the browser lands on it.
-        Returns (ok, thread_uuid). Updates last_verified_ts on success.
-        """
-        state = self.load_state()
-        rec = state.get(key)
-        if not rec or not rec.get("thread_uuid"):
-            return False, None
-        uuid = rec["thread_uuid"]
-        if not self.navigate_to_uuid(agent, uuid):
-            self.log_event("keepalive_check_failed",
-                           {"key": key, "agent": agent, "thread_uuid": uuid,
-                            "reason": "navigate_failed"})
-            return False, uuid
-        cur = self.current_thread_uuid(agent)
-        if cur == uuid:
-            rec["last_verified_ts"] = _ts()
-            rec["consecutive_failures"] = 0
-            state[key] = rec
-            self.save_state(state)
-            self.log_event("keepalive_check_ok",
-                           {"key": key, "agent": agent, "thread_uuid": uuid})
-            return True, uuid
-        self.log_event("keepalive_check_failed",
-                       {"key": key, "agent": agent, "thread_uuid": uuid,
-                        "reason": "url_mismatch", "current": cur})
-        return False, uuid
-
-    def ensure(self, key, agent, name=None):
-        """Ensure the thread exists and is reachable; create if needed.
-
-        Returns (ok, thread_uuid). Mirrors the job-dispatch.py reuse-or-
-        create flow: try stored UUID first, fall back to creation, then
-        capture the new UUID from the browser URL.
-        """
-        state = self.load_state()
-        rec = state.get(key)
-        if rec and rec.get("thread_uuid"):
-            ok, uuid = self.check(key, agent)
-            if ok:
-                return True, uuid
-            # Stored thread is dead — fall through to recreate.
-            print(f"keepalive: stored thread for {key} unreachable, "
-                  f"recreating", file=sys.stderr)
-        # Create a fresh sidechat.
-        if not self.create_sidechat(agent, name=name):
-            self.log_event("keepalive_create_failed",
-                           {"key": key, "agent": agent})
-            return False, None
-        # Capture the new thread UUID from the browser URL.
-        uuid = None
-        if not self.dry_run:
-            for _ in range(15):
-                time.sleep(1)
-                uuid = self.current_thread_uuid(agent)
-                if uuid:
-                    break
-        else:
-            uuid = "dry-run-uuid"
-        if not uuid:
-            self.log_event("keepalive_create_failed",
-                           {"key": key, "agent": agent,
-                            "reason": "uuid_capture_failed"})
-            return False, None
-        now = _ts()
-        state[key] = {
-            "thread_uuid": uuid,
-            "agent": agent,
-            "created_ts": now,
-            "last_ping_ts": 0,
-            "last_verified_ts": now,
-            "last_activity_ts": now,
-            "ping_count": 0,
-            "consecutive_failures": 0,
-        }
-        self.save_state(state)
-        self.log_event("keepalive_created",
-                       {"key": key, "agent": agent, "thread_uuid": uuid})
-        return True, uuid
-
-    def ping(self, key, agent, message=None):
-        """Send a keepalive ping into the thread.
-
-        Navigates to the thread first (cheap no-op if already there),
-        then sends the ping message. Updates last_ping_ts / ping_count.
-        """
-        state = self.load_state()
-        rec = state.get(key)
-        if not rec or not rec.get("thread_uuid"):
-            return False
-        uuid = rec["thread_uuid"]
-        if not self.navigate_to_uuid(agent, uuid):
-            return False
-        msg = (message or "[keepalive:{key}] tick {ts} {uuid}").format(
-            key=key, ts=_utcnow(), uuid=uuid)
-        if not self.send_message(agent, msg):
-            self.log_event("keepalive_ping_failed",
-                           {"key": key, "agent": agent, "thread_uuid": uuid})
-            return False
-        now = _ts()
-        rec["last_ping_ts"] = now
-        rec["last_activity_ts"] = now
-        rec["ping_count"] = rec.get("ping_count", 0) + 1
-        state[key] = rec
-        self.save_state(state)
-        self.log_event("keepalive_ping",
-                       {"key": key, "agent": agent, "thread_uuid": uuid,
-                        "ping_count": rec["ping_count"]})
-        return True
-
-    def note_activity(self, key):
-        """Record external activity (e.g. a job just sent to the thread)
-        so the idle timer doesn't ping unnecessarily."""
-        state = self.load_state()
-        rec = state.get(key)
-        if rec:
-            rec["last_activity_ts"] = _ts()
-            state[key] = rec
-            self.save_state(state)
-
-    def tick(self, key, agent, idle_threshold_s=3600, max_retries=3,
-             ping_message=None):
-        """One keepalive tick for a registered thread.
-
-        - Ensures the thread exists (recreate if dead).
-        - Pings only if idle longer than idle_threshold_s.
-        - After max_retries consecutive failures, forces recreation.
-        Returns a status string: ok | pinged | recreated | failed.
-        """
-        state = self.load_state()
-        rec = state.get(key)
-
-        if not rec or not rec.get("thread_uuid"):
-            ok, _ = self.ensure(key, agent)
-            return "recreated" if ok else "failed"
-
-        failures = rec.get("consecutive_failures", 0)
-        if failures >= max_retries:
-            # Force recreation: drop the dead UUID and re-ensure.
-            self.log_event("keepalive_force_recreate",
-                           {"key": key, "agent": agent,
-                            "failures": failures,
-                            "old_uuid": rec.get("thread_uuid")})
-            rec["thread_uuid"] = None
-            state[key] = rec
-            self.save_state(state)
-            ok, _ = self.ensure(key, agent)
-            return "recreated" if ok else "failed"
-
-        ok, _ = self.check(key, agent)
-        if not ok:
-            state = self.load_state()
-            rec = state.get(key, {})
-            rec["consecutive_failures"] = failures + 1
-            state[key] = rec
-            self.save_state(state)
-            return "failed"
-
-        now = _ts()
-        idle_for = now - rec.get("last_activity_ts", 0)
-        if idle_for >= idle_threshold_s:
-            if self.ping(key, agent, message=ping_message):
-                return "pinged"
-            return "failed"
-        return "ok"
-
-    def unregister(self, key):
-        """Remove a thread from keepalive (does not delete the sidechat)."""
-        state = self.load_state()
-        if key in state:
-            del state[key]
-            self.save_state(state)
-            self.log_event("keepalive_unregistered", {"key": key})
-            return True
-        return False
-
-    def status(self):
-        """Return the full registry with computed idle times."""
-        state = self.load_state()
-        now = _ts()
-        out = {}
-        for key, rec in state.items():
-            r = dict(rec)
-            r["idle_s"] = int(now - rec.get("last_activity_ts", now))
-            out[key] = r
-        return out

bin/kpi.py

@@ -1,707 +0,0 @@
-#!/usr/bin/env python3
-"""kpi.py — NetVM Fleet KPI, Spend Monitor & Runtime Preservation Engine.
-
-Monitors:
-  - Calls / DMs dispatched and verified (from dm-log.jsonl)
-  - Token quota spend & remaining (weekly limit % and extra tokens)
-  - Active subagent sessions and tmux muse workers
-  - Uptime vs actual problems fixed (Efficiency Index)
-  - Route health (WARP wireguard, CDP, tmux sockets)
-  - Runtime preservation advisories (guiding agents to offload work to tmux/subagents)
-"""
-
-from __future__ import annotations
-
-import argparse
-import json
-import os
-import re
-import subprocess
-import sys
-import time
-from dataclasses import asdict, dataclass
-from datetime import datetime, timezone
-from pathlib import Path
-from typing import Any, Dict, List, Optional
-
-REPO_ROOT = Path(__file__).resolve().parent.parent
-BIN_DIR = REPO_ROOT / "bin"
-DM_LOG = REPO_ROOT / "dm-log.jsonl"
-JOBS_DIR = REPO_ROOT / "jobs"
-SUBAGENTS_FILE = REPO_ROOT / "subagent-sessions.json"
-
-VALID_NODES = ["muse", "pip", "646", "opm", "dev", "def"]
-
-
-@dataclass
-class AgentKPI:
-    node: str
-    weekly_used_pct: Optional[int]
-    extra_tokens_remaining: str
-    is_blocked: bool
-    calls_sent: int
-    calls_verified: int
-    jobs_assigned: int
-    jobs_completed: int
-    subagents_active: int
-    tmux_workers_active: int
-    uptime_hours: float
-    route_status: str
-    efficiency_index: float
-    efficiency_rating: str
-    preservation_advisory: str
-
-    def to_dict(self) -> Dict[str, Any]:
-        return asdict(self)
-
-
-def get_agent_dm_metrics(node: str, window_hours: Optional[float] = None) -> Dict[str, int]:
-    """Calculate outbound messages, sends, and verified deliveries from dm-log.jsonl."""
-    if not DM_LOG.exists():
-        return {"sent": 0, "verified": 0, "total_events": 0}
-
-    cutoff = None
-    if window_hours:
-        cutoff = datetime.now(timezone.utc).timestamp() - (window_hours * 3600)
-
-    sent_ids = set()
-    verified_ids = set()
-    total_events = 0
-
-    try:
-        with open(DM_LOG, "r", encoding="utf-8") as f:
-            for line in f:
-                line = line.strip()
-                if not line:
-                    continue
-                try:
-                    entry = json.loads(line)
-                except Exception:
-                    continue
-
-                if entry.get("agent") != node:
-                    continue
-
-                if cutoff:
-                    ts = entry.get("ts")
-                    if ts:
-                        try:
-                            dt = datetime.fromisoformat(ts.replace("Z", "+00:00"))
-                            if dt.timestamp() < cutoff:
-                                continue
-                        except Exception:
-                            pass
-
-                total_events += 1
-                mid = entry.get("id")
-                etype = entry.get("type")
-                if etype in ("send_start", "send_done"):
-                    if mid:
-                        sent_ids.add(mid)
-                elif etype == "verified":
-                    if mid:
-                        verified_ids.add(mid)
-    except Exception:
-        pass
-
-    return {
-        "sent": len(sent_ids),
-        "verified": len(verified_ids),
-        "total_events": total_events,
-    }
-
-
-def get_agent_job_metrics(node: str) -> Dict[str, int]:
-    """Calculate total jobs assigned and completed for an agent."""
-    assigned = 0
-    completed = 0
-
-    if not JOBS_DIR.exists():
-        return {"assigned": 0, "completed": 0}
-
-    try:
-        for p in JOBS_DIR.glob("*.json"):
-            try:
-                with open(p, "r", encoding="utf-8") as f:
-                    data = json.load(f)
-                    if data.get("agent") == node:
-                        assigned += 1
-                        # If output or status has result
-                        if data.get("status") == "completed" or data.get("result"):
-                            completed += 1
-            except Exception:
-                continue
-    except Exception:
-        pass
-
-    return {"assigned": assigned, "completed": completed}
-
-
-def get_agent_subagent_count(node: str) -> int:
-    """Get active subagent sessions for a node from subagent-sessions.json."""
-    if not SUBAGENTS_FILE.exists():
-        return 0
-    try:
-        with open(SUBAGENTS_FILE, "r", encoding="utf-8") as f:
-            data = json.load(f)
-            if not isinstance(data, dict):
-                return 0
-            return sum(1 for s in data.values() if s.get("parent") == node and s.get("status") == "active")
-    except Exception:
-        return 0
-
-
-def get_agent_tmux_workers(node: str) -> List[str]:
-    """Get running tmux sessions for an agent (shared and netns socket)."""
-    sessions = []
-    # 1. Per-node socket
-    sock = f"/tmp/tmux-{node}.sock"
-    if os.path.exists(sock):
-        try:
-            r = subprocess.run(["tmux", "-S", sock, "list-sessions", "-F", "#{session_name}"], capture_output=True, text=True, timeout=2)
-            if r.returncode == 0 and r.stdout.strip():
-                sessions.extend(line.strip() for line in r.stdout.splitlines() if line.strip())
-        except Exception:
-            pass
-
-    # 2. Shared socket filtering sessions containing node name
-    shared_sock = "/tmp/tmux-muse.sock"
-    if os.path.exists(shared_sock):
-        try:
-            r = subprocess.run(["tmux", "-S", shared_sock, "list-sessions", "-F", "#{session_name}"], capture_output=True, text=True, timeout=2)
-            if r.returncode == 0 and r.stdout.strip():
-                for s in r.stdout.splitlines():
-                    s = s.strip()
-                    if s and (node in s or s.startswith(f"{node}-") or s == "swarm-worker"):
-                        if s not in sessions:
-                            sessions.append(s)
-        except Exception:
-            pass
-
-    return sessions
-
-
-def get_agent_uptime_hours(node: str) -> float:
-    """Calculate browser process uptime in hours."""
-    try:
-        # Search for chromium process matching user-data-dir or node
-        cmd = ["pgrep", "-f", f"chrome-box launch {node}"]
-        r = subprocess.run(cmd, capture_output=True, text=True, timeout=2)
-        pids = r.stdout.strip().split()
-        if not pids:
-            cmd = ["pgrep", "-f", f"profiles/{node}"]
-            r = subprocess.run(cmd, capture_output=True, text=True, timeout=2)
-            pids = r.stdout.strip().split()
-
-        if pids:
-            pid = pids[0]
-            # Read /proc/<pid>/stat starttime
-            stat_path = Path(f"/proc/{pid}/stat")
-            if stat_path.exists():
-                stat_content = stat_path.read_text().split()
-                # field 22 is starttime (in clock ticks after boot)
-                start_ticks = int(stat_content[21])
-                clk_tck = os.sysconf(os.sysconf_names["SC_CLK_TCK"])
-                with open("/proc/uptime", "r") as f:
-                    uptime_sec = float(f.read().split()[0])
-                process_age_sec = uptime_sec - (start_ticks / clk_tck)
-                return round(max(0.0, process_age_sec / 3600.0), 2)
-    except Exception:
-        pass
-    return 0.0
-
-
-def check_node_routes(node: str) -> str:
-    """Check connectivity route for a node (netns + CDP)."""
-    # 1. Check netns
-    netns_path = Path(f"/var/run/netns/warp-{node}")
-    if not netns_path.exists():
-        return "NO_NETNS"
-
-    # 2. Check CDP page connection
-    try:
-        try:
-            from approvals import get_node_pages
-        except ImportError:
-            sys.path.insert(0, str(BIN_DIR))
-            from approvals import get_node_pages
-        pages = get_node_pages(node, timeout=2.0)
-        if pages:
-            return "ONLINE"
-    except Exception:
-        pass
-
-    return "DEGRADED"
-
-
-
-def calculate_efficiency(
-    jobs_done: int,
-    subagents_active: int,
-    tmux_workers: int,
-    calls_verified: int,
-    weekly_used_pct: Optional[int],
-    uptime_hours: float,
-) -> tuple[float, str]:
-    """
-    Composite efficiency score.
-    Higher is better: measures actual work produced (jobs + subagents + workers + verified comms)
-    relative to quota burned and uptime elapsed.
-    """
-    work_units = (jobs_done * 5.0) + (subagents_active * 3.0) + (tmux_workers * 4.0) + (calls_verified * 0.5)
-    burn_cost = max(1.0, (weekly_used_pct or 10) * 0.2)
-
-    # Base index
-    index = round(work_units / burn_cost, 2)
-
-    # Classify
-    if uptime_hours > 2.0 and subagents_active == 0 and tmux_workers == 0 and jobs_done == 0:
-        rating = "VANITY_IDLE"
-    elif index >= 3.0:
-        rating = "HIGH_EFFICIENCY"
-    elif index >= 1.0:
-        rating = "PRODUCTIVE"
-    elif index >= 0.4:
-        rating = "MODERATE"
-    else:
-        rating = "LOW_EFFICIENCY"
-
-    return index, rating
-
-
-def generate_preservation_advisory(
-    node: str,
-    weekly_used_pct: Optional[int],
-    extra_tokens_remaining: str,
-    subagents_active: int,
-    tmux_workers: int,
-    rating: str,
-) -> str:
-    """Generate prescriptive runtime preservation instructions for the agent."""
-    tips = []
-
-    pct = weekly_used_pct or 0
-    if pct >= 95 or "0 tokens left" in extra_tokens_remaining:
-        return "CRITICAL: Quota exhausted. Do NOT send chat messages. Salvage via 'box onboard start <new_node> --for %s'." % node
-
-    if pct >= 70:
-        tips.append("Quota > 70%%: Cease prose chatter; offload tasks to background tmux workers.")
-
-    if subagents_active == 0 and tmux_workers == 0:
-        tips.append("Spawn subagents with isolated context ('box subagent spawn') or tmux muse workers.")
-
-    if rating in ("VANITY_IDLE", "LOW_EFFICIENCY"):
-        tips.append("Uptime without worker execution drains quota. Mandate: split goals into executable jobs.")
-
-    if not tips:
-        tips.append("Runtime healthy. Maintain worker-first execution strategy.")
-
-    return " ".join(tips)
-
-
-def get_agent_kpi(node: str, usage_cache: Optional[Dict[str, Any]] = None) -> AgentKPI:
-    """Collect comprehensive KPI metrics for a single NetVM node."""
-    # 1. Quota & Usage
-    usage = usage_cache
-    if usage is None:
-        try:
-            try:
-                import invite
-            except ImportError:
-                sys.path.insert(0, str(BIN_DIR))
-                import invite
-            u = invite.get_usage(node)
-            if isinstance(u, dict) and u.get("ok"):
-                usage = u
-        except Exception:
-            pass
-    if usage is None:
-        usage = {}
-
-    weekly_pct = usage.get("weekly_used_pct")
-    extra_left = usage.get("additional_left") or usage.get("extra_tokens_remaining") or "Unknown"
-    is_blocked = bool(usage.get("is_blocked")) or (weekly_pct is not None and weekly_pct >= 100 and "0 tokens left" in extra_left)
-
-    # 2. Activity metrics
-    dm_metrics = get_agent_dm_metrics(node)
-    job_metrics = get_agent_job_metrics(node)
-    subagents = get_agent_subagent_count(node)
-    tmux_sessions = get_agent_tmux_workers(node)
-    uptime = get_agent_uptime_hours(node)
-
-    route_status = check_node_routes(node)
-
-    # 3. Efficiency
-    eff_idx, eff_rating = calculate_efficiency(
-        jobs_done=job_metrics["completed"],
-        subagents_active=subagents,
-        tmux_workers=len(tmux_sessions),
-        calls_verified=dm_metrics["verified"],
-        weekly_used_pct=weekly_pct,
-        uptime_hours=uptime,
-    )
-
-    # 4. Advisory
-    advisory = generate_preservation_advisory(
-        node=node,
-        weekly_used_pct=weekly_pct,
-        extra_tokens_remaining=extra_left,
-        subagents_active=subagents,
-        tmux_workers=len(tmux_sessions),
-        rating=eff_rating,
-    )
-
-    return AgentKPI(
-        node=node,
-        weekly_used_pct=weekly_pct,
-        extra_tokens_remaining=extra_left,
-        is_blocked=is_blocked,
-        calls_sent=dm_metrics["sent"],
-        calls_verified=dm_metrics["verified"],
-        jobs_assigned=job_metrics["assigned"],
-        jobs_completed=job_metrics["completed"],
-        subagents_active=subagents,
-        tmux_workers_active=len(tmux_sessions),
-        uptime_hours=uptime,
-        route_status=route_status,
-        efficiency_index=eff_idx,
-        efficiency_rating=eff_rating,
-        preservation_advisory=advisory,
-    )
-
-
-def fleet_kpi(nodes: Optional[List[str]] = None) -> Dict[str, AgentKPI]:
-    """Collect KPI metrics across all fleet agents."""
-    target_nodes = nodes or VALID_NODES
-
-    # Fetch usage in bulk
-    usage_map = {}
-    try:
-        import invite
-        raw_usage = invite.fleet_usage(target_nodes)
-        if isinstance(raw_usage, dict):
-            usage_map = raw_usage
-    except Exception:
-        pass
-
-    results = {}
-    for n in target_nodes:
-        results[n] = get_agent_kpi(n, usage_cache=usage_map.get(n))
-    return results
-
-
-def get_live_advisory_block(node: str) -> str:
-    """Generate Markdown prompt envelope block ready for job injection."""
-    kpi = get_agent_kpi(node)
-    quota_str = f"{kpi.weekly_used_pct}% weekly limit used" if kpi.weekly_used_pct is not None else "quota active"
-    tokens_str = kpi.extra_tokens_remaining
-
-    lines = [
-        "---- BOX PERFORMANCE & RUNTIME ADVISORY ----",
-        f"AGENT: @{kpi.node} | QUOTA: {quota_str} ({tokens_str}) | UPTIME: {kpi.uptime_hours}h",
-        f"WORK UNITS: {kpi.jobs_completed} jobs finished | {kpi.subagents_active} subagents | {kpi.tmux_workers_active} tmux workers",
-        f"EFFICIENCY: {kpi.efficiency_rating} (Index: {kpi.efficiency_index}) | ROUTES: {kpi.route_status}",
-        f"RUNTIME MANDATE: {kpi.preservation_advisory}",
-        "Offload long operations to subagents or tmux muse workers to maximize problem-fixing per token.",
-    ]
-    return "\n".join(lines)
-
-
-NODE_SIDECHATS = {
-    "646": "646 tasks",
-    "opm": "heartbeat",
-    "pip": "646-pip-coord",
-    "dev": "dev-coord",
-    "def": "def-coord",
-    "muse": "646-muse-coord",
-}
-
-
-def find_pending_work_for_node(node: str) -> Optional[Dict[str, Any]]:
-    """Find assigned pending job or swarm slot for an agent node."""
-    # 1. Look for node-specific auto-work jobs
-    if JOBS_DIR.exists():
-        candidates = sorted(list(JOBS_DIR.glob(f"auto-work-{node}-*.json")) + list(JOBS_DIR.glob(f"{node}-*.json")))
-        for c in candidates:
-            try:
-                with open(c, "r", encoding="utf-8") as f:
-                    data = json.load(f)
-                job_agent = data.get("agent")
-                if job_agent and job_agent != node:
-                    continue
-                job_name = c.stem
-                return {
-                    "type": "job",
-                    "name": job_name,
-                    "path": str(c),
-                    "cmd": f"{sys.executable} {BIN_DIR}/job-dispatch.py {job_name}",
-                }
-            except Exception:
-                continue
-
-    # 2. Check pending swarm slots
-    try:
-        from swarm_worker.poller import find_pending_slots
-        slots = find_pending_slots()
-        if slots:
-            slot = slots[0]
-            sw_id = slot.get("swarm_id", "swarm")
-            idx = slot.get("slot_index", 0)
-            return {
-                "type": "swarm",
-                "name": f"swarm-{sw_id}-s{idx}",
-                "path": None,
-                "cmd": f"{sys.executable} {BIN_DIR}/swarm_worker/daemon.py",
-            }
-    except Exception:
-        pass
-
-    return None
-
-
-def auto_spawn_workers(nodes: Optional[List[str]] = None, dry_run: bool = False) -> List[Dict[str, Any]]:
-    """Reconcile idle agents and auto-spawn background tmux workers to execute pending work."""
-    target_nodes = nodes or VALID_NODES
-    results = []
-
-    for node in target_nodes:
-        # Check active tmux workers for this node
-        active_tmux = len(get_agent_tmux_workers(node))
-        if active_tmux > 0:
-            results.append({
-                "node": node,
-                "action": "skip",
-                "reason": f"Active tmux worker already running ({active_tmux})",
-            })
-            continue
-
-        # Check work availability
-        work = find_pending_work_for_node(node)
-        if not work:
-            results.append({
-                "node": node,
-                "action": "idle",
-                "reason": "No pending jobs or swarm slots",
-            })
-            continue
-
-        session_label = f"worker-{work['name'][:18]}"
-        cmd_to_run = f"{work['cmd']} > /tmp/tmux-{node}-{session_label}.log 2>&1"
-
-        if dry_run:
-            results.append({
-                "node": node,
-                "action": "would_spawn",
-                "session": session_label,
-                "work_type": work["type"],
-                "work_name": work["name"],
-                "command": work["cmd"],
-            })
-            continue
-
-        # Execute spawn
-        spawn_res = spawn_tmux_worker(node, session_label, cmd_to_run)
-        if spawn_res.get("ok"):
-            # Send sidechat notification
-            try:
-                from invite_handler import send_loopback_notice
-                chat = NODE_SIDECHATS.get(node, "646 tasks")
-                msg = f"[BOX-AUTO-WORKER] Spawned background tmux worker '{session_label}' executing {work['type']} ({work['name']}). Logs at /tmp/tmux-{node}-{session_label}.log"
-                send_loopback_notice(recipient=node, target=chat, message=msg)
-            except Exception:
-                pass
-
-            results.append({
-                "node": node,
-                "action": "spawned",
-                "session": session_label,
-                "work_type": work["type"],
-                "work_name": work["name"],
-                "command": work["cmd"],
-            })
-        else:
-            results.append({
-                "node": node,
-                "action": "error",
-                "error": spawn_res.get("error", "Unknown spawn error"),
-            })
-
-    return results
-
-
-def spawn_tmux_worker(node: str, session: str, command: str) -> Dict[str, Any]:
-    """Spawn an autonomous tmux worker session on the agent's netns or shared socket."""
-    # Ensure session name is prefixed
-    clean_session = f"{node}-{session}" if not session.startswith(f"{node}-") else session
-
-    try:
-        from subagent_tracker import register_session
-    except ImportError:
-        sys.path.insert(0, str(BIN_DIR))
-        from subagent_tracker import register_session
-
-    muse_tmux = BIN_DIR / "muse-tmux.py"
-    if not muse_tmux.exists():
-        return {"ok": False, "error": "muse-tmux.py not found"}
-
-    # Execute via muse-tmux.py
-    cmd = [
-        sys.executable,
-        str(muse_tmux),
-        "new",
-        clean_session,
-        "--node",
-        node,
-        "--command",
-        command,
-    ]
-    res = subprocess.run(cmd, capture_output=True, text=True, timeout=10)
-    if res.returncode != 0:
-        # Fallback to shared socket
-        cmd_shared = [
-            sys.executable,
-            str(muse_tmux),
-            "new",
-            clean_session,
-            "--command",
-            command,
-        ]
-        res = subprocess.run(cmd_shared, capture_output=True, text=True, timeout=10)
-        if res.returncode != 0:
-            return {"ok": False, "error": res.stderr.strip() or res.stdout.strip()}
-
-    # Register in subagent tracker
-    sid = f"tmux-{clean_session}-{int(time.time())}"
-    register_session(parent=node, session_id=sid, title=f"tmux-worker-{clean_session}", prompt=command)
-
-    return {
-        "ok": True,
-        "node": node,
-        "session": clean_session,
-        "session_id": sid,
-        "command": command,
-        "message": f"Spawned tmux worker '{clean_session}' for @{node}. Running in background.",
-    }
-
-
-def main():
-    parser = argparse.ArgumentParser(description="NetVM Fleet KPI, Spend Monitor & Runtime Preservation Engine")
-    subparsers = parser.add_subparsers(dest="command")
-
-    p_status = subparsers.add_parser("status", help="Show fleet KPI metrics table")
-    p_status.add_argument("--node", choices=VALID_NODES, help="Filter by node")
-    p_status.add_argument("--json", action="store_true", help="Emit JSON output")
-
-    p_report = subparsers.add_parser("report", help="Detailed KPI report for a specific node")
-    p_report.add_argument("node", choices=VALID_NODES, help="Target node")
-    p_report.add_argument("--json", action="store_true")
-
-    p_routes = subparsers.add_parser("routes", help="Verify network and CDP routes across nodes")
-    p_routes.add_argument("--json", action="store_true")
-
-    p_block = subparsers.add_parser("prompt-block", help="Generate live prompt envelope block for node")
-    p_block.add_argument("node", choices=VALID_NODES, help="Target node")
-
-    p_spawn = subparsers.add_parser("spawn-worker", help="Spawn autonomous background tmux worker session")
-    p_spawn.add_argument("node", choices=VALID_NODES, help="Agent node")
-    p_spawn.add_argument("session", help="Session label")
-    p_spawn.add_argument("worker_command", help="Command to execute inside worker")
-    p_autospawn = subparsers.add_parser("auto-spawn", help="Auto-spawn background tmux workers for idle nodes with pending work")
-    p_autospawn.add_argument("--node", choices=VALID_NODES, default=None, help="Filter by node")
-    p_autospawn.add_argument("--dry-run", action="store_true", help="Report what would be spawned without executing")
-    p_autospawn.add_argument("--json", action="store_true")
-
-    args = parser.parse_args()
-
-    if args.command in (None, "status"):
-        nodes = [args.node] if getattr(args, "node", None) else VALID_NODES
-        kpis = fleet_kpi(nodes)
-        if getattr(args, "json", False):
-            print(json.dumps({k: v.to_dict() for k, v in kpis.items()}, indent=2))
-            return
-
-        print("\n=== NETVM FLEET KPI & RUNTIME PRESERVATION DASHBOARD ===\n")
-        header = f"{'NODE':<6} {'QUOTA':<10} {'CALLS':<12} {'JOBS':<10} {'SUBAGENTS':<11} {'TMUX':<6} {'UPTIME':<8} {'ROUTE':<9} {'EFFICIENCY':<15}"
-        sep = f"{'────':<6} {'─────────':<10} {'───────────':<12} {'─────────':<10} {'──────────':<11} {'────':<6} {'──────':<8} {'───────':<9} {'──────────────':<15}"
-        print(header)
-        print(sep)
-        for n in nodes:
-            k = kpis.get(n)
-            if not k:
-                continue
-            q_str = f"{k.weekly_used_pct}%" if k.weekly_used_pct is not None else "Active"
-            c_str = f"{k.calls_sent} ({k.calls_verified}v)"
-            j_str = f"{k.jobs_completed}/{k.jobs_assigned}"
-            sub_str = str(k.subagents_active)
-            tmux_str = str(k.tmux_workers_active)
-            up_str = f"{k.uptime_hours}h"
-            print(f"{k.node:<6} {q_str:<10} {c_str:<12} {j_str:<10} {sub_str:<11} {tmux_str:<6} {up_str:<8} {k.route_status:<9} {k.efficiency_rating:<15}")
-        print("\nRun 'box kpi report <node>' for prescriptive runtime preservation advisories.\n")
-
-    elif args.command == "report":
-        kpi = get_agent_kpi(args.node)
-        if args.json:
-            print(json.dumps(kpi.to_dict(), indent=2))
-            return
-        print(f"\n=== KPI & RUNTIME REPORT: @{kpi.node.upper()} ===")
-        print(f"  Weekly Quota:       {kpi.weekly_used_pct}% used")
-        print(f"  Extra Tokens:       {kpi.extra_tokens_remaining}")
-        print(f"  Blocked Status:     {'YES (LIMIT REACHED)' if kpi.is_blocked else 'NO (HEALTHY)'}")
-        print(f"  Messages / Calls:   {kpi.calls_sent} sent ({kpi.calls_verified} verified delivered)")
-        print(f"  Jobs Dispatched:    {kpi.jobs_completed} completed / {kpi.jobs_assigned} assigned")
-        print(f"  Active Subagents:   {kpi.subagents_active}")
-        print(f"  Active Tmux Workers:{kpi.tmux_workers_active}")
-        print(f"  Process Uptime:     {kpi.uptime_hours} hours")
-        print(f"  Route Health:       {kpi.route_status}")
-        print(f"  Efficiency Index:   {kpi.efficiency_index} ({kpi.efficiency_rating})")
-        print(f"\n  [RUNTIME PRESERVATION ADVISORY]\n  {kpi.preservation_advisory}\n")
-
-    elif args.command == "routes":
-        routes = {n: check_node_routes(n) for n in VALID_NODES}
-        if args.json:
-            print(json.dumps(routes, indent=2))
-        else:
-            print("\n=== NETVM ROUTE HEALTH ===")
-            for n, st in routes.items():
-                print(f"  @{n:<6} : {st}")
-            print()
-
-    elif args.command == "prompt-block":
-        print(get_live_advisory_block(args.node))
-
-    elif args.command == "spawn-worker":
-        res = spawn_tmux_worker(args.node, args.session, args.worker_command)
-        if args.json:
-            print(json.dumps(res, indent=2))
-        else:
-            if res.get("ok"):
-                print(f"✔ {res.get('message')}")
-            else:
-                print(f"✘ Failed to spawn worker: {res.get('error')}", file=sys.stderr)
-                sys.exit(1)
-
-    elif args.command == "auto-spawn":
-        nodes = [args.node] if getattr(args, "node", None) else None
-        results = auto_spawn_workers(nodes=nodes, dry_run=args.dry_run)
-        if args.json:
-            print(json.dumps(results, indent=2))
-            return
-        print(f"\n=== AUTO-SPAWN WORKER RECONCILIATION {'(DRY-RUN)' if args.dry_run else ''} ===")
-        for r in results:
-            n = r.get("node")
-            act = r.get("action")
-            if act == "spawned":
-                print(f"  ✔ @{n:<5} : SPAWNED session '{r.get('session')}' ({r.get('work_type')}: {r.get('work_name')})")
-            elif act == "would_spawn":
-                print(f"  ? @{n:<5} : WOULD SPAWN session '{r.get('session')}' ({r.get('work_type')}: {r.get('work_name')})")
-            elif act == "skip":
-                print(f"  - @{n:<5} : SKIP ({r.get('reason')})")
-            elif act == "idle":
-                print(f"  - @{n:<5} : IDLE ({r.get('reason')})")
-            elif act == "error":
-                print(f"  ✘ @{n:<5} : ERROR ({r.get('error')})")
-        print()
-
-
-if __name__ == "__main__":
-    main()

bin/lookup

@@ -1 +0,0 @@
-/home/super/Projects/NetVM/bin/docs-lookup.py

bin/lookup_engine.py

@@ -1,200 +0,0 @@
-#!/usr/bin/env python3
-"""lookup_engine.py — Shared Zero-Downtime Hot-Reloading Pattern & Schema Engine.
-
-Provides authoritative runtime access to lookup_internal/ databases for
-daemons (response-harvester, self_main_loop, job-dispatch), CLI commands,
-and agents.
-
-Features:
-- Dynamic mtime-based zero-downtime hot reloading of compiled regex patterns.
-- Pre-flight soft validation of outbound agent sentences and work orders.
-- Direct helper functions for core protocol regexes (RESULT, VERB, TOOL, etc.).
-"""
-
-import json
-import os
-import re
-import sys
-from pathlib import Path
-from typing import Any, Dict, List, Optional, Tuple
-
-NETVM_ROOT = Path(__file__).resolve().parent.parent
-LOOKUP_INTERNAL = NETVM_ROOT / "lookup_internal"
-if not LOOKUP_INTERNAL.exists() and (NETVM_ROOT / "docs_internal").exists():
-    LOOKUP_INTERNAL = NETVM_ROOT / "docs_internal"
-
-# In-memory cache structures with modification timestamps
-_CACHE_MTIMES: Dict[str, float] = {}
-_RAW_CACHE: Dict[str, Any] = {}
-_COMPILED_PATTERNS: Dict[str, re.Pattern] = {}
-
-# Fallback hardcoded regexes in case files are missing or unreadable
-_FALLBACK_RESULT_RE = re.compile(r"\[RESULT\s+([A-Za-z0-9_/-]+)\]\s*(.*?)(?=\[RESULT\s|\Z)", re.S)
-_FALLBACK_VERB_RE = re.compile(r"\[(ACK|CLAIM|RESULT|DECLINE|NO-ACTION)\s+([A-Za-z0-9_/-]+)\]")
-_FALLBACK_TOOL_RE = re.compile(r"\[(TOOL|EXEC|DM)\s+(?:([a-zA-Z0-9_.-]+)\s+)?(\{([^{}]|\{[^{}]*\})*\})\]", re.S)
-_FALLBACK_CONTRACT_FOOTER = (
-    "Reply: [ACK id] seen | [CLAIM id] mine | "
-    "[RESULT id] done | [DECLINE id] | [NO-ACTION id]."
-)
-
-
-def load_lookup_json(filename: str) -> Dict[str, Any]:
-    """Load JSON from lookup_internal/ with mtime-based caching."""
-    target_path = LOOKUP_INTERNAL / filename
-    if not target_path.is_file():
-        return {}
-
-    try:
-        current_mtime = os.path.getmtime(target_path)
-    except OSError:
-        return _RAW_CACHE.get(filename, {})
-
-    if filename in _RAW_CACHE and _CACHE_MTIMES.get(filename) == current_mtime:
-        return _RAW_CACHE[filename]
-
-    try:
-        with open(target_path, "r", encoding="utf-8") as f:
-            data = json.load(f)
-        _RAW_CACHE[filename] = data
-        _CACHE_MTIMES[filename] = current_mtime
-        return data
-    except Exception as e:
-        print(f"[lookup_engine] Warning: Error reading {target_path}: {e}", file=sys.stderr)
-        return _RAW_CACHE.get(filename, {})
-
-
-def _refresh_compiled_patterns_if_needed():
-    """Checks regex_patterns.json mtime and recompiles if changed."""
-    global _COMPILED_PATTERNS
-    data = load_lookup_json("regex_patterns.json")
-    patterns_data = data.get("patterns", {})
-
-    target_path = LOOKUP_INTERNAL / "regex_patterns.json"
-    current_mtime = _CACHE_MTIMES.get("regex_patterns.json", 0.0)
-    compiled_mtime = _CACHE_MTIMES.get("_compiled_patterns_mtime", 0.0)
-
-    if current_mtime == compiled_mtime and _COMPILED_PATTERNS:
-        return
-
-    new_compiled = {}
-    for key, entry in patterns_data.items():
-        raw_pat = entry.get("pattern", "")
-        flag_names = entry.get("flags", [])
-        flags = 0
-        for fn in flag_names:
-            if hasattr(re, fn):
-                flags |= getattr(re, fn)
-        try:
-            new_compiled[key] = re.compile(raw_pat, flags)
-        except Exception as e:
-            print(f"[lookup_engine] Warning: Failed to compile pattern '{key}': {e}", file=sys.stderr)
-
-    _COMPILED_PATTERNS = new_compiled
-    _CACHE_MTIMES["_compiled_patterns_mtime"] = current_mtime
-
-
-def get_compiled_pattern(name: str) -> Optional[re.Pattern]:
-    """Retrieve a compiled pattern by name, hot-reloading if the database was modified."""
-    _refresh_compiled_patterns_if_needed()
-    return _COMPILED_PATTERNS.get(name)
-
-
-def get_all_compiled_patterns() -> Dict[str, re.Pattern]:
-    """Retrieve all compiled patterns with automatic hot-reloading."""
-    _refresh_compiled_patterns_if_needed()
-    return dict(_COMPILED_PATTERNS)
-
-
-def get_result_regex() -> re.Pattern:
-    """Return the canonical [RESULT ...] regex."""
-    p = get_compiled_pattern("result")
-    return p if p is not None else _FALLBACK_RESULT_RE
-
-
-def get_verb_regex() -> re.Pattern:
-    """Return the canonical [VERB ...] regex (ACK|CLAIM|RESULT|DECLINE|NO-ACTION)."""
-    p = get_compiled_pattern("verb")
-    return p if p is not None else _FALLBACK_VERB_RE
-
-
-def get_tool_regex() -> re.Pattern:
-    """Return the canonical [TOOL ...] regex."""
-    p = get_compiled_pattern("tool_call")
-    return p if p is not None else _FALLBACK_TOOL_RE
-
-
-def get_contract_footer() -> str:
-    """Return standard contract footer string."""
-    struct_data = load_lookup_json("sentence_structure.json")
-    cf = struct_data.get("structures", {}).get("contract_footer", {})
-    return cf.get("example") or _FALLBACK_CONTRACT_FOOTER
-
-
-def parse_agent_utterance(text: str) -> List[Dict[str, Any]]:
-    """Pass text through all registered patterns and extract matched tokens."""
-    _refresh_compiled_patterns_if_needed()
-    patterns_data = load_lookup_json("regex_patterns.json").get("patterns", {})
-    matches = []
-
-    for key, compiled in _COMPILED_PATTERNS.items():
-        m = compiled.search(text)
-        if m:
-            entry = patterns_data.get(key, {})
-            matches.append({
-                "pattern_key": key,
-                "pattern_name": entry.get("name", key),
-                "matched_text": m.group(0),
-                "named_groups": m.groupdict(),
-                "span": m.span()
-            })
-    return matches
-
-
-def validate_outbound_sentence(text: str) -> Tuple[bool, Optional[str], Optional[str]]:
-    """Pre-flight check for outbound messages sent via CLI.
-
-    Returns:
-        (is_valid, matched_kind, warning_or_hint)
-    """
-    cleaned = text.strip()
-    # If message starts with bracketed protocol marker
-    if cleaned.startswith("["):
-        marker = cleaned.split("]")[0] + "]"
-        upper_marker = marker.upper()
-
-        if upper_marker.startswith("[WO:") or upper_marker.startswith("[WORKORDER:"):
-            wo_pat = get_compiled_pattern("work_order")
-            if wo_pat and not wo_pat.search(cleaned):
-                hint = (
-                    "Notice: Message starts with a Work Order marker but does not match canonical structure.\n"
-                    "  Expected format: [WO:<id>] [from <sender>] <title> — <body>\n"
-                    "  Example: [WO:7fce46e0] [from super] Audit endpoints — Check GET /api/stats\n"
-                    "  Hint: Query 'box lookup sentence work_order' for full spec."
-                )
-                return False, "work_order", hint
-            return True, "work_order", None
-
-        if upper_marker.startswith("[ACK:") or upper_marker.startswith("[ACK "):
-            ack_pat = get_compiled_pattern("ack")
-            verb_pat = get_compiled_pattern("verb")
-            if (ack_pat and not ack_pat.search(cleaned)) and (verb_pat and not verb_pat.search(cleaned)):
-                hint = (
-                    "Notice: Message looks like an ACK but deviates from standard syntax.\n"
-                    "  Expected format: [ACK:<id>] [from <sender>] or [ACK <id>]\n"
-                    "  Example: [ACK:7fce46e0] [from 646]"
-                )
-                return False, "ack", hint
-            return True, "ack", None
-
-        if upper_marker.startswith("[RESULT"):
-            res_pat = get_result_regex()
-            if not res_pat.search(cleaned):
-                hint = (
-                    "Notice: Message starts with [RESULT] but deviates from standard syntax.\n"
-                    "  Expected format: [RESULT <job_id>] <status> <summary>\n"
-                    "  Example: [RESULT 7fce46e0] OK Task completed successfully"
-                )
-                return False, "result", hint
-            return True, "result", None
-
-    return True, "plain_message", None

bin/main-chat-watchdog.py

@@ -1,194 +0,0 @@
-#!/usr/bin/env python3
-"""main-chat-watchdog.py — enforce the sidechat-first DM policy.
-
-Tails dm-log.jsonl (read-only) and flags any DM actually SENT to main chat
-without an explicit allow_main_chat marker, plus any placement_mismatch /
-placement_failed events (message landed in the wrong chat). Distinguishes:
-
-  VIOLATION : type=sent, target=main, no tags.allow_main_chat  -> policy broken (P0)
-  AUTHORIZED: type=sent, target=main, tags.allow_main_chat=true -> explicit opt-in
-  BLOCKED   : type=main_chat_blocked                             -> policy working
-  MISMATCH  : type=placement_mismatch                             -> placed in wrong chat (P0)
-  PFAILED   : type=placement_failed                               -> placement hard-failed (P0)
-
-State: byte-offset watermark in main-chat-watchdog.state (handles log
-rotation via inode check). First run starts at EOF (no historical backfill —
-old entries predate the allow_main_chat audit marker).
-
-Violations are appended to logs/main-chat-violations.jsonl and printed to
-stdout (journal). Exit 0 = clean, 1 = violations/P0s found, 2 = error.
-
-Read-only: never modifies dm-log.jsonl.
-"""
-
-import json
-import os
-import sys
-from datetime import datetime, timezone
-
-BASE = "/home/super/Projects/NetVM"
-DM_LOG = os.path.join(BASE, "dm-log.jsonl")
-STATE_FILE = os.path.join(BASE, "main-chat-watchdog.state")
-VIOLATIONS_LOG = os.path.join(BASE, "logs", "main-chat-violations.jsonl")
-
-
-def utcnow():
-    return datetime.now(timezone.utc).isoformat()
-
-
-def load_state():
-    try:
-        with open(STATE_FILE) as f:
-            return json.load(f)
-    except (FileNotFoundError, json.JSONDecodeError, ValueError):
-        return {}
-
-
-def save_state(state):
-    tmp = STATE_FILE + ".tmp"
-    with open(tmp, "w") as f:
-        json.dump(state, f)
-    os.replace(tmp, STATE_FILE)
-
-
-def main():
-    try:
-        st = os.stat(DM_LOG)
-    except FileNotFoundError:
-        print(f"watchdog ERROR: {DM_LOG} not found", file=sys.stderr)
-        return 2
-
-    state = load_state()
-    # First run (or rotation): start at EOF, don't backfill history that
-    # predates the allow_main_chat audit marker.
-    if state.get("inode") != st.st_ino:
-        offset = st.st_size
-        if state:
-            print(f"watchdog: log rotated or first run (inode {st.st_ino}), "
-                  f"starting at EOF offset {offset}")
-    else:
-        offset = min(state.get("offset", 0), st.st_size)
-
-    violations = []   # P0: gate bypass (sent to main, no opt-in)
-    mismatches = []   # P0: placement_mismatch / placement_failed
-    blocked = 0
-    authorized = 0
-    scanned = 0
-    malformed = 0
-
-    try:
-        with open(DM_LOG, "r", encoding="utf-8", errors="replace") as f:
-            f.seek(offset)
-            for line in f:
-                line = line.strip()
-                if not line:
-                    continue
-                scanned += 1
-                try:
-                    ev = json.loads(line)
-                except json.JSONDecodeError:
-                    malformed += 1
-                    continue
-                etype = ev.get("type")
-                if etype == "main_chat_blocked":
-                    # Policy working: the dm.py gate refused a main send.
-                    blocked += 1
-                elif etype == "placement_mismatch":
-                    # P0: verify-time URL check found the browser parked in a
-                    # different chat than the intended target. The send
-                    # completed but landed in the wrong place. Schema has two
-                    # variants: sidechat ("expected_uuid") and main-drift
-                    # ("expected": "main").
-                    mismatches.append({
-                        "ts": utcnow(),
-                        "kind": "placement_mismatch",
-                        "severity": "P0",
-                        "dm_id": ev.get("id"),
-                        "agent": ev.get("agent"),
-                        "to": ev.get("to"),
-                        "target": ev.get("target"),
-                        "expected_uuid": ev.get("expected_uuid") or ev.get("expected"),
-                        "actual_uuid": ev.get("actual_uuid"),
-                        "attempt": ev.get("attempt"),
-                        "event_ts": ev.get("ts"),
-                    })
-                elif etype == "placement_failed":
-                    # P0: the authoritative post-send placement check failed
-                    # hard (sender exited 1, send NOT marked verified).
-                    d = ev.get("detail") or {}
-                    mismatches.append({
-                        "ts": utcnow(),
-                        "kind": "placement_failed",
-                        "severity": "P0",
-                        "dm_id": ev.get("id"),
-                        "agent": ev.get("agent"),
-                        "to": ev.get("to"),
-                        "target": ev.get("target"),
-                        "reason": ev.get("reason") or d.get("reason"),
-                        "expected_uuid": ev.get("expected_uuid") or d.get("expected_uuid"),
-                        "actual_url": ev.get("actual_url") or d.get("actual_url"),
-                        "loop_attempt": ev.get("loop_attempt"),
-                        "event_ts": ev.get("ts"),
-                    })
-                elif etype == "sent" and ev.get("target") == "main":
-                    tags = ev.get("tags") or {}
-                    if tags.get("allow_main_chat"):
-                        authorized += 1
-                    else:
-                        violations.append({
-                            "ts": utcnow(),
-                            "kind": "main_chat_send",
-                            "severity": "P0",
-                            "dm_id": ev.get("id"),
-                            "agent": ev.get("agent"),
-                            "to": ev.get("to"),
-                            "target": "main",
-                            "msg_preview": (ev.get("msg") or "")[:120],
-                            "event_ts": ev.get("ts"),
-                        })
-            new_offset = f.tell()
-    except OSError as e:
-        print(f"watchdog ERROR reading log: {e}", file=sys.stderr)
-        return 2
-
-    save_state({"offset": new_offset, "inode": st.st_ino})
-
-    findings = violations + mismatches
-    summary = (f"watchdog: scanned={scanned} blocked={blocked} "
-               f"authorized_main={authorized} "
-               f"violations={len(violations)} "
-               f"placement_mismatches={len(mismatches)} "
-               f"malformed={malformed}")
-    print(summary)
-
-    if findings:
-        try:
-            os.makedirs(os.path.dirname(VIOLATIONS_LOG), exist_ok=True)
-            with open(VIOLATIONS_LOG, "a", encoding="utf-8") as vf:
-                for v in findings:
-                    vf.write(json.dumps(v) + "\n")
-        except OSError as e:
-            print(f"watchdog ERROR writing violations log: {e}", file=sys.stderr)
-            return 2
-        for v in violations:
-            print(f"VIOLATION main-chat send without opt-in: "
-                  f"id={v['dm_id']} agent={v['agent']} to={v['to']} "
-                  f"at={v['event_ts']} preview={v['msg_preview']!r}")
-        for m in mismatches:
-            if m["kind"] == "placement_mismatch":
-                print(f"P0 PLACEMENT_MISMATCH message landed in wrong chat: "
-                      f"id={m['dm_id']} agent={m['agent']} to={m['to']} "
-                      f"target={m['target']} expected={m['expected_uuid']} "
-                      f"actual={m['actual_uuid']} at={m['event_ts']}")
-            else:
-                print(f"P0 PLACEMENT_FAILED placement check failed: "
-                      f"id={m['dm_id']} agent={m['agent']} to={m['to']} "
-                      f"target={m['target']} reason={m['reason']} "
-                      f"expected={m['expected_uuid']} "
-                      f"actual_url={m['actual_url']} at={m['event_ts']}")
-        return 1
-    return 0
-
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/meta-ac-snapshot.py

@@ -1,205 +0,0 @@
-#!/usr/bin/env python3
-"""Meta Accounts Center change-detection harness.
-
-Captures a structural snapshot of the accountscenter.meta.com auth flow
-via CDP inside a NetVM netns, diffs against the stored baseline.
-
-Outcomes:
-  PASS    - matches baseline (or first run establishes it)
-  CHANGED - structural diff detected; needs human review, baseline untouched
-  FAIL    - automation itself broke (browser/CDP/network error)
-
-Usage:
-  meta-ac-snapshot.py [--node NAME] [--promote] [--snapshot-dir DIR]
-
-  --node      NetVM node to run in (default: phone)
-  --promote   after human review, promote the latest snapshot to baseline
-  --snapshot-dir  where snapshots live (default: ~/Projects/NetVM/snapshots/meta-ac)
-"""
-import argparse, base64, datetime, json, os, subprocess, sys, time
-import urllib.parse, urllib.request
-
-CDP_PORT = 19744
-
-def log(*a):
-    print(*a, flush=True)
-
-def ns_exec(node, cmd):
-    return subprocess.run(
-        ["sudo", "-n", "ip", "netns", "exec", f"warp-{node}"] + cmd,
-        capture_output=True, text=True)
-
-def norm_url(u):
-    """Strip query/fragment — nonces change every visit."""
-    p = urllib.parse.urlparse(u)
-    return f"{p.scheme}://{p.host}{p.path}" if hasattr(p, 'host') else f"{p.scheme}://{p.hostname}{p.path}"
-
-def main():
-    ap = argparse.ArgumentParser()
-    ap.add_argument("--node", default="phone")
-    ap.add_argument("--promote", action="store_true")
-    ap.add_argument("--snapshot-dir", default=os.path.expanduser(
-        "~/Projects/NetVM/snapshots/meta-ac"))
-    args = ap.parse_args()
-    os.makedirs(args.snapshot_dir, exist_ok=True)
-    baseline_path = os.path.join(args.snapshot_dir, "baseline.json")
-
-    if args.promote:
-        snaps = sorted(f for f in os.listdir(args.snapshot_dir)
-                       if f.startswith("snap-") and f.endswith(".json"))
-        if not snaps:
-            log("no snapshots to promote"); return 2
-        latest = os.path.join(args.snapshot_dir, snaps[-1])
-        data = json.load(open(latest))
-        data["promoted_at"] = datetime.datetime.now(datetime.timezone.utc).isoformat()
-        json.dump(data, open(baseline_path, "w"), indent=2)
-        log(f"promoted {snaps[-1]} -> baseline.json")
-        return 0
-
-    profile_dir = "/tmp/meta-ac-snap-profile"
-    subprocess.run(["rm", "-rf", profile_dir])
-    os.makedirs(profile_dir, exist_ok=True)
-
-    def http(path):
-        with urllib.request.urlopen(
-                f"http://127.0.0.1:{CDP_PORT}{path}", timeout=5) as r:
-            return json.loads(r.read())
-
-    # launch chromium inside the netns via a wrapper script
-    wrapper = "/tmp/meta-ac-snap-run.py"
-    open(wrapper, "w").write(WRAPPER_SRC)
-    log(f"launching chromium in warp-{args.node} (CDP {CDP_PORT})...")
-    proc = subprocess.Popen(
-        ["sudo", "-n", "ip", "netns", "exec", f"warp-{args.node}",
-         "python3", wrapper, str(CDP_PORT), profile_dir],
-        stdout=subprocess.PIPE, stderr=subprocess.STDOUT, text=True)
-    try:
-        out, _ = proc.communicate(timeout=120)
-    except subprocess.TimeoutExpired:
-        proc.kill(); log("FAIL: harness timed out"); return 1
-    print(out)
-    # wrapper prints SNAPSHOT_JSON=<json> on success
-    snap = None
-    for line in out.splitlines():
-        if line.startswith("SNAPSHOT_JSON="):
-            snap = json.loads(line[len("SNAPSHOT_JSON="):])
-    if not snap:
-        log("FAIL: no snapshot captured"); return 1
-
-    snap["node"] = args.node
-    snap["captured_at"] = datetime.datetime.now(datetime.timezone.utc).isoformat()
-
-    # egress ip for context
-    try:
-        r = ns_exec(args.node, ["curl", "-s", "--max-time", "8",
-                                "https://api.ipify.org"])
-        snap["egress_ip"] = r.stdout.strip()
-    except Exception:
-        snap["egress_ip"] = "unknown"
-
-    ts = datetime.datetime.now(datetime.timezone.utc).strftime("%Y%m%d-%H%M%S")
-    snap_path = os.path.join(args.snapshot_dir, f"snap-{ts}.json")
-    json.dump(snap, open(snap_path, "w"), indent=2)
-    log(f"snapshot saved: {snap_path}")
-
-    if not os.path.exists(baseline_path):
-        json.dump(snap, open(baseline_path, "w"), indent=2)
-        log("PASS: baseline established (first run)")
-        return 0
-
-    baseline = json.load(open(baseline_path))
-    diffs = diff_snapshots(baseline, snap)
-    if not diffs:
-        log("PASS: matches baseline")
-        return 0
-    log("CHANGED: structural diff detected (baseline untouched):")
-    for d in diffs:
-        log(f"  - {d}")
-    log("review with: diff baseline.json snap-<ts>.json")
-    log("promote after review with: --promote")
-    return 3
-
-def diff_snapshots(base, snap):
-    diffs = []
-    b_chain = [norm_url(u) for u in base.get("redirect_chain", [])]
-    s_chain = [norm_url(u) for u in snap.get("redirect_chain", [])]
-    if b_chain != s_chain:
-        diffs.append(f"redirect_chain changed: {b_chain} -> {s_chain}")
-    for key in ("forms", "inputs", "buttons"):
-        b = sorted(base.get("dom_markers", {}).get(key, []))
-        s = sorted(snap.get("dom_markers", {}).get(key, []))
-        if b != s:
-            added = [x for x in s if x not in b]
-            removed = [x for x in b if x not in s]
-            diffs.append(f"dom_markers.{key}: added={added} removed={removed}")
-    if base.get("final_title") != snap.get("final_title"):
-        diffs.append(f"final_title: {base.get('final_title')!r} -> {snap.get('final_title')!r}")
-    return diffs
-
-WRAPPER_SRC = '''
-import json, subprocess, sys, time, os, urllib.request, base64
-CDP_PORT = int(sys.argv[1])
-PROFILE_DIR = sys.argv[2]
-def http(path):
-    with urllib.request.urlopen(f"http://127.0.0.1:{CDP_PORT}{path}", timeout=5) as r:
-        return json.loads(r.read())
-logf = open("/tmp/meta-ac-snap-chrome.log", "w")
-proc = subprocess.Popen(["chromium", "--headless=new", "--disable-gpu", "--no-sandbox",
-    "--disable-dev-shm-usage", f"--user-data-dir={PROFILE_DIR}",
-    f"--remote-debugging-port={CDP_PORT}", "--remote-allow-origins=*", "about:blank"],
-    stdout=logf, stderr=subprocess.STDOUT)
-try:
-    for i in range(30):
-        try:
-            ver = http("/json/version")
-            if "webSocketDebuggerUrl" in ver: break
-        except Exception: pass
-        time.sleep(1)
-    else:
-        print("FAIL: CDP never came up"); sys.exit(1)
-    import websocket
-    bws = websocket.create_connection(ver["webSocketDebuggerUrl"], timeout=20)
-    bws.send(json.dumps({"id": 1, "method": "Target.createTarget",
-        "params": {"url": "https://accountscenter.meta.com"}}))
-    target_id = json.loads(bws.recv())["result"]["targetId"]
-    bws.close()
-    # redirect chain: seed with the navigation target (we always start
-    # there), then poll for where Meta sends us. Seeding fixes the race
-    # where a fast redirect is missed by the poll interval.
-    START_URL = "https://accountscenter.meta.com/"
-    chain, seen = [START_URL], {START_URL}
-    for _ in range(24):
-        time.sleep(2)
-        for t in http("/json/list"):
-            if t.get("id") == target_id or "meta.com" in t.get("url", ""):
-                u = t["url"]
-                if u not in seen:
-                    seen.add(u); chain.append(u)
-                title = t.get("title", "")
-                break
-    # dom markers from the final tab
-    tab_ws = None
-    for t in http("/json/list"):
-        if t.get("id") == target_id or "meta.com" in t.get("url", ""):
-            tab_ws = t["webSocketDebuggerUrl"]; final_url = t["url"]; break
-    ws = websocket.create_connection(tab_ws, timeout=20)
-    js = """JSON.stringify({
-        forms: [...document.forms].map(f => f.id || f.name || '(anon)'),
-        inputs: [...document.querySelectorAll('input')].map(i => i.name || i.type || '(anon)'),
-        buttons: [...document.querySelectorAll('button, [role=button]')].map(b => (b.innerText||'').trim()).filter(Boolean)
-    })"""
-    ws.send(json.dumps({"id": 1, "method": "Runtime.evaluate",
-        "params": {"expression": js, "returnByValue": True}}))
-    markers = json.loads(json.loads(ws.recv())["result"]["result"]["value"])
-    # dedupe buttons, keep order
-    markers["buttons"] = list(dict.fromkeys(markers["buttons"]))
-    ws.close()
-    snap = {"redirect_chain": chain, "final_url": final_url,
-            "final_title": title, "dom_markers": markers}
-    print("SNAPSHOT_JSON=" + json.dumps(snap))
-finally:
-    proc.terminate()
-'''
-
-if __name__ == "__main__":
-    sys.exit(main())

bin/meta-acct.py

@@ -1,218 +0,0 @@
-#!/usr/bin/env python3
-"""Meta Accounts Center read API (via authenticated browser session).
-
-Usage:
-  meta-acct.py list-linked <agent>       # linked profiles in this Meta Account
-  meta-acct.py security-status <agent>   # login & recovery / security checkup summary
-  meta-acct.py login-activity <agent>    # "Where you're logged in" sessions
-
-The agent's browser must have an active Meta session (phone/email OTP login).
-Reads NODES.md for the CDP port. Outputs JSON.
-
-Scope: Accounts Center linkage/security surface only. No writes.
-"""
-import json, sys, time, urllib.request, os
-
-NETVM = os.environ.get("NETVM_DIR")
-if not NETVM:
-    script_dir = os.path.dirname(os.path.abspath(__file__))
-    parent = os.path.dirname(script_dir)
-    if os.path.exists(os.path.join(parent, "NODES.md")):
-        NETVM = parent
-    else:
-        NETVM = "/home/super/Projects/NetVM"
-
-AC_BASE = "https://accountscenter.meta.com"
-
-def cdp_port(agent):
-    for line in open(os.path.join(NETVM, "NODES.md")):
-        line = line.strip()
-        if not line.startswith("|"):
-            continue
-        cells = [c.strip() for c in line.strip("|").split("|")]
-        if len(cells) >= 4 and cells[0] == agent:
-            return int(cells[3])
-    raise SystemExit(f"meta-acct: unknown agent '{agent}' (not in NODES.md)")
-
-def cdp_get(port, path, method="GET", timeout=10):
-    req = urllib.request.Request(f"http://127.0.0.1:{port}{path}", method=method)
-    with urllib.request.urlopen(req, timeout=timeout) as r:
-        return json.loads(r.read())
-
-def get_ac_tab(port):
-    tabs = cdp_get(port, "/json/list")
-    for t in tabs:
-        if "accountscenter.meta.com" in t.get("url", "") and t.get("type") == "page":
-            return t
-    # create one
-    return cdp_get(port, f"/json/new?{AC_BASE}/", method="PUT")
-
-def evaluate(port, ws_url, js, timeout=15):
-    import websocket
-    ws = websocket.create_connection(ws_url, timeout=timeout)
-    try:
-        ws.send(json.dumps({"id": 1, "method": "Page.navigate",
-                            "params": {"url": js[0]}}))
-        ws.recv()
-        time.sleep(4)
-        ws.send(json.dumps({"id": 2, "method": "Runtime.evaluate",
-            "params": {"expression": js[1], "returnByValue": True}}))
-        resp = json.loads(ws.recv())
-        return json.loads(resp["result"]["result"]["value"])
-    finally:
-        ws.close()
-
-def cmd_list_linked(port):
-    tab = get_ac_tab(port)
-    data = evaluate(port, tab["webSocketDebuggerUrl"], (
-        f"{AC_BASE}/account_overview/",
-        """JSON.stringify((() => {
-          const h2 = [...document.querySelectorAll('h2')]
-            .find(e=>e.innerText.trim()==='Profiles');
-          // Full section is 3 levels up (H2 > DIV > DIV > MAIN)
-          const container = h2?.parentElement?.parentElement?.parentElement;
-          return {
-            email: (document.body.innerText.match(
-              /[\\w.+-]+@[\\w-]+\\.[\\w.]+/)||[])[0]||null,
-            section: container?.innerText?.slice(0,1000) || null
-          };
-        })())"""))
-    # Parse: lines after "Profiles" are name/type pairs until "Add more",
-    # then "AI and devices" section follows
-    profiles = []
-    section = data.get("section", "")
-    if section:
-        lines = [l.strip() for l in section.split("\n") if l.strip()]
-        try:
-            start = lines.index("Profiles") + 1
-        except ValueError:
-            start = 0
-        i = start
-        while i < len(lines):
-            if lines[i] == "Add more":
-                i += 1
-                continue
-            if lines[i] == "AI and devices":
-                # Next line is the device/service name
-                if i + 1 < len(lines):
-                    profiles.append({"name": lines[i+1], "type": "ai_device"})
-                break
-            # Expect name + type pair
-            if i + 1 < len(lines) and lines[i+1] not in (
-                    "Add more", "AI and devices", "Profiles"):
-                profiles.append({"name": lines[i],
-                                 "type": lines[i+1].lower()})
-                i += 2
-            else:
-                i += 1
-    return {"email": data.get("email"), "profiles": profiles}
-
-def cmd_security_status(port):
-    tab = get_ac_tab(port)
-    data = evaluate(port, tab["webSocketDebuggerUrl"], (
-        f"{AC_BASE}/password_and_security/",
-        """JSON.stringify({
-          checkup: document.body.innerText.match(
-            /Meta Security Checkup\\s*(\\d+) recommended actions/)?.[1] || null,
-          sections: [...document.querySelectorAll('h2')]
-            .map(e=>e.innerText.trim()).slice(0,10),
-          body: document.body.innerText.slice(0,2000)
-        })"""))
-    return {"security_checkup_actions": data.get("checkup"),
-            "sections": data.get("sections")}
-
-def cmd_login_activity(port):
-    tab = get_ac_tab(port)
-    # "Where you're logged in" is on the password_and_security page
-    data = evaluate(port, tab["webSocketDebuggerUrl"], (
-        f"{AC_BASE}/password_and_security/",
-        """JSON.stringify({
-          body: document.body.innerText.slice(0,4000)
-        })"""))
-    body = data.get("body", "")
-    # Extract the "Where you're logged in" section
-    idx = body.find("Where you're logged in")
-    section = body[idx:idx+1500] if idx >= 0 else ""
-    return {"where_logged_in": section}
-
-def cmd_link_instagram(agent, port):
-    """Automate Meta Accounts Center linking flow and second-click OAuth handoff."""
-    import websocket
-    tab = get_ac_tab(port)
-    ws = websocket.create_connection(tab["webSocketDebuggerUrl"], timeout=15)
-    try:
-        # Step 1: Navigate to manage accounts
-        ws.send(json.dumps({"id": 1, "method": "Page.navigate", "params": {"url": f"{AC_BASE}/manage/"}}))
-        ws.recv()
-        time.sleep(3)
-
-        # Step 2: Look for 'Add profiles and devices' button or check if already linked
-        expr_find_add = """(() => {
-            const btns = [...document.querySelectorAll("div[role='button'], button, a")];
-            const addBtn = btns.find(b => /add profiles|add accounts/i.test((b.innerText||"").trim()));
-            if (addBtn) {
-                addBtn.click();
-                return {status: "clicked_add", text: addBtn.innerText};
-            }
-            return {status: "no_add_button"};
-        })()"""
-        ws.send(json.dumps({"id": 2, "method": "Runtime.evaluate", "params": {"expression": expr_find_add, "returnByValue": True}}))
-        res2 = json.loads(ws.recv()).get("result", {}).get("result", {}).get("value", {})
-        time.sleep(3)
-
-        # Step 3: Check dialog / popup for Instagram option or FXCAL handoff
-        expr_handle_dialog = """(() => {
-            const btns = [...document.querySelectorAll("div[role='button'], button, a")];
-            // Check for Add Instagram button or Continue/Confirm
-            const igBtn = btns.find(b => /instagram/i.test((b.innerText||"").trim()) && /add|connect/i.test((b.innerText||"").trim()));
-            if (igBtn) {
-                igBtn.click();
-                return {status: "clicked_instagram_option"};
-            }
-            const confirmBtn = btns.find(b => /continue|confirm|yes, finish/i.test((b.innerText||"").trim()));
-            if (confirmBtn) {
-                confirmBtn.click();
-                return {status: "clicked_confirm", text: confirmBtn.innerText};
-            }
-            return {status: "dialog_scanned", current_url: window.location.href};
-        })()"""
-        ws.send(json.dumps({"id": 3, "method": "Runtime.evaluate", "params": {"expression": expr_handle_dialog, "returnByValue": True}}))
-        res3 = json.loads(ws.recv()).get("result", {}).get("result", {}).get("value", {})
-        time.sleep(2)
-
-        return {
-            "status": "success",
-            "step1_add": res2,
-            "step2_dialog": res3,
-            "current_url": tab.get("url")
-        }
-    finally:
-        ws.close()
-
-def main():
-    if len(sys.argv) < 3:
-        print(__doc__.strip().split("\n")[0])
-        print("Usage: meta-acct.py <list-linked|security-status|login-activity|link-instagram> <agent>")
-        sys.exit(2)
-    cmd, agent = sys.argv[1], sys.argv[2]
-    port = cdp_port(agent)
-    try:
-        if cmd == "list-linked":
-            out = cmd_list_linked(port)
-        elif cmd == "security-status":
-            out = cmd_security_status(port)
-        elif cmd == "login-activity":
-            out = cmd_login_activity(port)
-        elif cmd == "link-instagram":
-            out = cmd_link_instagram(agent, port)
-        else:
-            raise SystemExit(f"meta-acct: unknown command '{cmd}'")
-    except Exception as e:
-        print(json.dumps({"error": str(e), "agent": agent}))
-        sys.exit(1)
-    out["agent"] = agent
-    out["checked_at"] = int(time.time())
-    print(json.dumps(out, indent=2))
-
-if __name__ == "__main__":
-    main()