reap-stale
Skill2 starsChanged 5 months agoApache-2.0View on GitHub
Sweep the SDLC queue for issues stuck in status:in-progress whose heartbeat has gone silent (default >30 min) and reclaim them back to status:ready so other workers can pick them up. Trigger when the user says "reap stale issues", "clean up the queue", "release stuck claims", "/reap-stale", or notices an issue claimed for too long. Safe to run on a cron via the schedule skill. This is the fast (30-min) complement to .github/workflows/issue-hygiene.yml's slower 36h reaper.
In this repository9 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- context-syncSkill
- deploy-sanitySkill
- engine-adapter-scaffoldSkill
- file-issueSkill
- ir-schema-guardSkill
- issue-triageSkill
- pick-issueSkill
- tdd-cycleSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 25
Run from the main checkout (absolute path or cd to repo root).
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/reap-stale — Janitor for the SDLC pipeline
Sweep the GitHub issue queue and reclaim any abandoned status:in-progress
claim. An abandoned claim = a sticky claim comment whose Last heartbeat is
older than heartbeat_stale_after_seconds (default 1800s).
This is the safety net: if a worker agent crashes, runs out of context, or
quietly drops, its issue would otherwise be locked. The reaper releases it back
to status:ready so the queue keeps moving. It sits alongside
.github/workflows/issue-hygiene.yml, which reaps on a slower 36h cron — this
one enforces the 30-min heartbeat lease for tight parallel-agent turnaround.
pick-issue.sh already runs the reaper before every claim, so you rarely need
this manually. Use it when:
- You see an issue stuck
status:in-progressfor a while and want it freed now. - You're cron'ing it via
/schedulefor proactive cleanup. - You're debugging a stuck queue.
Run from the main checkout (absolute path or cd to repo root).
Invocation
# Real run — actually reclaims stale issues:
.claude/scripts/sdlc/reap-stale.sh
# Dry run — reports what WOULD be reclaimed, changes nothing:
.claude/scripts/sdlc/reap-stale.sh --dry-runOutput, one line per in-progress issue:
reap: 3 in-progress issue(s); threshold=1800s
#42: alive (heartbeat 120s old)
#43: STALE (heartbeat 2400s old, threshold 1800s) — reaping
#44: no claim comment found — releasing as orphanedWhat it does on reap
- Deletes the stale
claim-v1sticky comment(s). - Swaps
status:in-progress→status:ready, dropsagent:claimed. - Posts a
♻️ Reclaimedcomment noting the heartbeat age. - Removes local
.claude/sdlc/state/claim-<num>.json.
Threshold
Set in .claude/sdlc/config.json under heartbeat_stale_after_seconds
(default 1800s = 30 min). Don't lower it below ~5 min — reaper false-positives
(stealing an issue from a live-but-slow worker) hurt worse than false-negatives.
Sibling skills: /pick-issue, /file-issue. Pipeline docs: .claude/sdlc/README.md.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

