feat(swarm-worker): dispatch prose swarm slots to ephemeral Muse subagents

- daemon: route non-shell slot tasks to a fresh subagent session on dev/def/muse
  via muse_hybrid; add bin/ to sys.path so muse_hybrid/prompt_envelope import
  under the tmux supervisor
- prompt_envelope: add wrap_subagent_task() - plain task assignment without
  [TOOL tmux/swarm/cron] meta tags (subagents refused those as relayed test traffic)
- box-ctl/reporter: swarm-attach accepts optional session-id, stored as
  slot.subagent_session_id
- executor: _looks_like_shell requires an existing executable (prose like
  'verify ...' no longer misread as shell)
- poller: pick up pending swarms as well as running
- response-harvester: monitor running slot subagent sessions from swarms.json,
  allow '/' in RESULT/VERB job ids, archive ephemeral threads on any verdict
  (OK or FAIL), reap slot sessions for terminal swarms

bin/prompt_envelope.py

@@ -119,3 +119,18 @@ def wrap(job_name, job_id, agent, target, rendered):
         bottom += f"Conclude with your [RESULT {job_id}] line reporting outcomes.\n"
     return top + rendered.strip() + "\n" + bottom
 
+
+def wrap_subagent_task(job_id, task_text):
+    """Format an authentic direct task assignment for a subagent worker without test-traffic meta tags."""
+    return (
+        f"Operator assignment for swarm slot {job_id}:\n\n"
+        f"Task:\n"
+        f"{task_text.strip()}\n\n"
+        f"Instructions:\n"
+        f"1. Carry out this task directly using your available tools.\n"
+        f"2. When finished, conclude your final response with your verdict line:\n"
+        f"[RESULT {job_id}] OK: <one-line summary of actions and outcome>\n"
+        f"(or [RESULT {job_id}] FAIL: <reason> if the task could not be completed)\n"
+    )
+
+