fix(dispatch): eliminate noisy SSH signatures from routine job prompts and activate real tool triggers

bin/job-dispatch.py

@@ -222,8 +222,10 @@ def send_dm(agent, target, message, dry_run=False, followup_tags=None,
     if HAS_RATE_LIMITER:
         rate_limit_wait(agent)
 
-    # Cryptographic attestation: sign message with local SSH key
+    # 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():
@@ -234,7 +236,6 @@ def send_dm(agent, target, message, dry_run=False, followup_tags=None,
                 )
                 if sign_res.returncode == 0 and "-----BEGIN SSH SIGNATURE-----" in sign_res.stdout:
                     signed_payload = sign_res.stdout.strip()
-                # Extract message id and register proof to crypt.muse-dev.online
                     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:

jobs/box-http-health.json

@@ -11,11 +11,12 @@
   },
   "name": "box-http-health",
   "on_failure": "alert",
-  "prompt_template": "Box HTTP health check.\nJob ID: {job_id}\nTime: {datetime}\n\nRun the Box HTTP checks on the VM:\n  ssh dev-operator-646@34.139.37.135 \"/srv/box/bin/box-health-check.sh http\"\n(or run /srv/box/bin/box-health-check.sh http via any VM access you have)\n\nAlso verify the new live Box Console endpoints:\n1. https://box.muse-dev.online/ (Front-Door Console loads, assets box.js and box.css return 200)\n2. https://box.muse-dev.online/api/box/fleet (Returns 200 with live fleet array)\n3. https://box.muse-dev.online/api/box/dm/log (Returns 200 with recent DM log entries)\n\nExpected: all OK. If any FAIL, investigate immediately (check board.service,\nCaddy, recent deploys) and include the failure lines in your reply.\n\nReply with OK or FAIL <one-line summary>.",
+  "prompt_template": "Box HTTP endpoint health check.\nJob ID: {job_id}\nTime: {datetime}\n\nRun the fleet health check:\n[TOOL health.check {}]\n\nCheck recent scheduled runs:\n[TOOL cron.runs {}]\n\nIf all endpoints and fleet nodes return green, reply with [RESULT {job_id}] OK.\nOtherwise reply with [RESULT {job_id}] FAIL <summary>.",
   "schedule": "*/15 * * * *",
   "sidechat": {
     "create": true,
-    "name_template": "box-http-health-{datetime}"
+    "name_template": "box-http-health-{datetime}",
+    "reuse_key": "box-http-health"
   },
   "timeout": 600
 }

jobs/box-service-health.json

@@ -1,7 +1,7 @@
 {
   "agent": "646",
   "chain_next": null,
-  "description": "Box service + data health — every 15 minutes (offset 7m)",
+  "description": "Box service + data health \u2014 every 15 minutes (offset 7m)",
   "followup": {
     "escalate": "opm",
     "expect_reply": true,
@@ -11,11 +11,12 @@
   },
   "name": "box-service-health",
   "on_failure": "alert",
-  "prompt_template": "Box service health check.\nJob ID: {job_id}\nTime: {datetime}\n\nRun the Box service and data checks on the VM:\n  ssh dev-operator-646@34.139.37.135 \"/srv/box/bin/box-health-check.sh services\"\n  ssh dev-operator-646@34.139.37.135 \"/srv/box/bin/box-health-check.sh data\"\n(or run /srv/box/bin/box-health-check.sh via any VM access you have)\n\nExpected: board.service active, caddy active, sweeper timer firing,\n/srv/box writable. If any FAIL, investigate and include failure lines.\n\nReply with [RESULT {job_id}] OK or [RESULT {job_id}] FAIL <one-line summary>.",
+  "prompt_template": "Box service and data health check.\nJob ID: {job_id}\nTime: {datetime}\n\nExecute the service check on the VM:\n[TOOL service.status {\"unit\": \"board.service\"}]\n\nCheck caddy and harvester timer status:\n[TOOL service.status {\"unit\": \"response-harvester.timer\"}]\n\nIf healthy, end with [RESULT {job_id}] OK.\nIf any service fails, reply with [RESULT {job_id}] FAIL <summary>.",
   "schedule": "7,22,37,52 * * * *",
   "sidechat": {
     "create": true,
-    "name_template": "box-service-health-{datetime}"
+    "name_template": "box-service-health-{datetime}",
+    "reuse_key": "box-service-health"
   },
   "timeout": 600
 }