The supervisor agent was confusingly named "factory" (same as the
project). Rename directory, script, log, lock, status, and escalation
files. Update all references across scripts and docs.
FACTORY_ROOT env var unchanged (refers to project root, not agent).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add unconditional worktree cleanup to factory supervisor:
- Remove review + dev worktrees older than 2h with no active agent
- Use git worktree remove --force instead of rm -rf
- Run git worktree prune every poll to clear dangling refs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
pgrep matched all claude processes including manual screen sessions.
Narrow to "claude -p" so only non-interactive (factory-spawned) processes
get reaped by the stale-process cleanup.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add matrix_send() to lib/env.sh and matrix_listener.sh daemon for
real-time notifications, threaded escalations, and human-in-the-loop
replies. All agents now notify via Matrix instead of openclaw.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When dev-agent exhausts CI fix budget, writes escalation marker.
Supervisor picks it up next poll and invokes claude -p to diagnose
(flaky test? rate limit? real bug?) and decide: fix, close PR, or
escalate to human.
PR #649 sat blocked for hours — CI passed, review approved, but
merge conflict from other PRs landing. Supervisor said 'all clear'.
Now checks mergeable=false before CI state, alerts on conflict.
- PROMPT.md references best-practices/ files instead of inlining all knowledge
- best-practices/{memory,disk,ci,dev-agent,git}.md — loaded on demand by claude
- All alerts go to claude -p. Claude decides what to fix and what to escalate.
- update-prompt.sh targets specific best-practices files for self-learning
