github
Skill0 starsChanged 27 days agoMITView on GitHub
Applies Conventional Commits and branch naming, opens PRs, and polls review bots and CI until the PR can merge. Use when the user commits, opens a PR or addresses review feedback. Not for reviewing a diff (review) or plain git mechanics (execution).
- Commits and pushes
174 other projects use a version of this.
Adopted by
Most recently taken up by
- br3eze-code/br3eze-codetheir own version, first seen Nov 16, 2026
- fworks-tech/agenthoodtheir own version, first seen Nov 15, 2026
- heymegabyte/claude-skillstheir own version, first seen Nov 15, 2026
- heymegabyte/agent-skillstheir own version, first seen Nov 15, 2026
- rodos4312/Meu-Agenttheir own version, first seen Nov 15, 2026
- mynameistito/skillstheir own version, first seen Nov 15, 2026
Used with
Projects with it also use taskflow-inbox-triage, BlueBubbles and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
AtomicBot-aiSkilltaskflow-inbox-triage
Also in 6 projects that use this
jiulingyunSkillBlueBubbles
为Clawdbot构建或更新BlueBubbles外部通道插件(扩展包,REST发送/探测,webhook入站)。
Also in 20 projects that use this
hkyutongSkillnode-connect
Diagnose OpenClaw node connection and pairing failures for Android, iOS, and macOS companion apps. Use when QR/setup code/manual connect fails, local Wi-Fi works but VPS/tailnet does not, or errors mention pairing required, unauthorized, bootstrap token invalid or expired, gateway.bind, gateway.remote.url, Tailscale, or plugins.entries.device-pair.config.publicUrl.
Also in 18 projects that use this
molt-botSkilldiscord-e2e
Test Discord with Convex-leased QA bots: check credentials without a Gateway, exercise owned messages, or prove OpenClaw replies, files, threads, reactions, streaming, and typing. Distinguish API and Gateway evidence from manual user interactions.
Also in 10 projects that use this
molt-botSkillopenclaw-repair-sweep
Run scoped OpenClaw issue/PR repair campaigns: coordinate workers, prove root causes, and land or close verified work under the requested authority.
Also in 10 projects that use this
clawdbotSkillopenclaw-performance
Performance work in OpenClaw: a slow or heavy flow (latency, CPU, memory, SQL, RPC, renders), writing or reviewing a perf PR, or adding a cache. Count the work, delete waste first, cache last.
Also in 7 projects that use this
- +2 more from molt-bot
Other versions174 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 174
- openclaw/openclaw★ 395ksame name
- moltbot/clawdbot★ 394ksame name
- steipete/warelay★ 394ksame name
- steipete/clawdis★ 394ksame name
- moltbot/moltbot★ 394ksame name
- clawdbot/clawdbot★ 394ksame name
- molt-bot/clawdbot★ 391ksame name
- NousResearch/hermes-agent★ 264ksame name
- spinabot/brigade★ 11ksame name
- the-open-agent/openagent★ 5.7ksame name
164 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository22 other files it gives its agents.
- AGENTS.mdAGENTS.md
- claude-arsenal/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- coverage-gapsSkill
- dep-upgradeSkill
- designSkill
- executionSkill
- explore-ideaSkill
- gate-checkSkill
- initSkill
- litcriticSkill
- mutmut-reportSkill
- pin-checkSkill
- pypi-releaseSkill
- python-bootstrapSkill
- queue-addSkill
- queue-nextSkill
- queue-statusSkill
- reviewSkill
- session-endSkill
- shipSkill
- specifySkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 0 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 157 edited versions we read keep these.
0
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 52
Before gh pr create, run the host repo's lint/format/test gate (make lint, npm run lint, …), since the review loop assumes CI was green at …
- Very specific: line 83
After gh pr create returns the PR number, start the loop with the rubric inline, so each tick carries its own instructions, and with …
“Common” is measured: kept by at least 6 in 10 of the 157 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
github
Apply commit and PR conventions, then run an automated review loop instead of asking the user to relay bot comments. After a PR opens, the loop watches what gates a merge — CI, review-bot reactions and comments, and merge conflicts — answers each comment, and tells the user when the PR is ready.
CANARY: github-loaded-2026-05-20-d436255c-54a7f770e04e4983
When to load
- The user asks to commit, open a PR, or amend a PR body.
- The user asks whether CI is done, what a review bot said, or to address the review.
- A
/loopcycle is checking PR state.
Judging whether a diff is correct belongs to review or execution; this skill owns
the mechanics of the review loop.
Commit conventions
Conventional Commits: <type>(scope): description. Types: feat, fix, refactor,
test, docs, chore, ci. Scope is the module or domain. Imperative voice, no
trailing period, first line ≤72 chars. Body after a blank line explains why.
End the commit with the Co-Authored-By: line the harness's git instructions supply,
copied verbatim; add no other model name to commits, PR bodies, prose or
examples, because the harness value is the one that stays current.
PR conventions
## Summary
<1-3 bullets>
## Test plan
- [ ] <verifiable check>Branches: feat/<short-description>, fix/<short-description>. Main branch is main.
For several PRs that merge in sequence, load
stacking before creating the second branch.
Pre-PR gate
Before gh pr create, run the host repo's lint/format/test gate (make lint,
npm run lint, …), since the review loop assumes CI was green at push time. If the
project has no lint target, say so, propose one, and proceed.
FAST="${CLAUDE_SKILL_DIR}/../init/assets/bin/fast_gate.sh"
bash "$FAST" # each review round: `preflight-gate` over the changed files
bash "$FAST" --full # before merge, only if no receipt or green CI covers the tree--full writes a receipt reused while the tree is unchanged. Review rounds run the
fast gate once, then push once, because every push is a CI run. Load
claude-arsenal:core:init § references/ci-minutes.md when CI minutes are a concern.
A green gate says the change does not break the repo, not that it is the change that
was asked for; run the independent review before opening the PR, and on the open PR
when its decision line calls for one: claude-arsenal:core:init § references/pre-pr-review.md.
The review loop
In a cloud session that offers PR activity events (subscribe_pr_activity is in
its tools), subscribe to the PR instead of starting /loop, then end the turn: CI
results, reviews, comments and conflicts each wake the session, so polling only spends
turns on what is already pushed. On each wake, run query_pr_state.py --pr <N>
--unresolved-only --trigger once and act on its state as listed below and per the
rubric in the loop command. Some changes send no event (a silent bot, a reaction, the
quiet window, mergeability), so whenever the script exits 1 in a state other than
ci_running, schedule one check-in rather than a loop. If the subscription result says
another agent already receives the events, run the loop instead. Everywhere else, run
the loop.
After gh pr create returns the PR number, start the loop with the rubric inline, so
each tick carries its own instructions, and with --unresolved-only, so handled
comments drop out:
/loop 90s python3 "${CLAUDE_SKILL_DIR}/scripts/query_pr_state.py" --pr <PR_NUMBER> --unresolved-only --trigger — if state is bot_commented, address per the rubric (agree → fix, then after the round's last fix run fast_gate.sh once and push once + reply "addressed in <sha>" via gh api repos/<owner>/<repo>/pulls/<PR_NUMBER>/comments/<id>/replies; disagree → reply with rationale on the same endpoint; ambiguous → reply asking for clarification + ping the user). Pair every fix or dismissal with a reply on the thread, since the reply is what --unresolved-only filters on next tick. If bot_skipped / bot_rate_limited / bot_absent, follow its decision: local-review none → continue; diff or full → run one review round per the pre-PR review protocol, then add --local-review-done. If conflicts, rebase onto (or merge) the base branch, resolve, and push. If ci_failed, fetch the failing job log, fix, commit and push, then reply on related comments. Stop only on ready_to_merge, merged, or closed (bot_approved still waits for the quiet window); then CronDelete <job-id> and hand back to the user./loop rounds 90s up to every 2 minutes. Stop early with CronDelete <job-id>
(CronList recovers the ID).
The script prints JSON and exits 0 when there is something to act on or the loop is
done, 1 while waiting, 2 on conflicts or ci_failed. Exit 2 with any other state
(authentication, repo not found) is permanent: stop and tell the user.
conflicts— rebase onto the base branch, resolve, push. When the parent PR of a stacked branch merged, userebase_stack.shfrom stacking.ci_failed—gh run view --log-failed <run-id>, fix, push, reply on related comments.ready_to_merge— tell the user "PR #N ready to merge". In a repo witharsenal/config.toml, readmerge-policyand the vendored protocol's completion step first: the host may already have answered whether an agent merges.
Load pr-review-loop when a state is unclear
(bot_eyeing that persists, a bot that never reviews), when configuring watched
bots or bot-triggers, or when judging a comment the inline rubric does not settle.
Projects Classic vs v2
Projects Classic breaks gh pr view --comments and gh pr edit --body. On first use
in a repo:
python3 "${CLAUDE_SKILL_DIR}/scripts/query_project_type.py" --write-claude-mdIt prints classic / v2 / none and records <!-- github-skill: projects=… -->
in CLAUDE.md so later sessions skip detection. When it says classic, load
projects-detection for the workarounds.
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.

