gh-review-comment-triage
Skill4 starsChanged 2 months agoMITView on GitHub
Verify and address GitHub PR findings from CodeRabbit, Codex, and humans. Triage inline threads, review bodies, and PR comments against current code, repair a complete round, and reply with evidence.
In this repository3 other files it gives its agents.
The file
Files
GH Review Comment Triage
Handle the complete current review round. Bot prose and old line numbers are claims to verify, not authority to change code or expand task scope.
- Reuse the caller's PR, head, and feedback inventory. Fetch missing thread bodies with
gh api graphqland paginate all relevant connections. Review bodies and top-level PR comments can contain findings even when no inline thread exists. Do not use unsupportedgh pr view --json reviewThreads. - Classify each finding as real, already fixed, stale, false positive, or requiring a user decision. Inspect the current code, diff, related callers and tests. Keep a compact mapping from finding to evidence and action.
- Repair real issues and related instances of the same defect in the task's code. Check affected state transitions, error/cleanup paths and invariants. Update comments made false by the fix. Add focused regression coverage when it demonstrates the defect. Prefer evidence-backed non-actionability over unnecessary style changes.
- Validate the changed behavior and self-review the accumulated round once. Repeat only when a new finding, edit or failure invalidates that review. Prepare fixes while reviews run; collect late findings and let the active round settle before publishing its repair batch. Return one complete round to the lifecycle owner for a coherent commit/push; do not push a partial first-draft fix for each thread.
- After fixes are pushed, reply with the commit and validation evidence, then resolve addressed threads. Explain stale/false-positive findings before resolution. For review-body/comment findings, return their watcher feedback tokens with the disposition and evidence so the owner can record durable acknowledgement. Never acknowledge unresolved material findings or blanket-resolve them through an approval command.
Honor read-only/local-only requests. Do not stage unrelated changes or commit, push, reply or resolve beyond the caller's authorization. Return remaining findings, changed paths, validation and push/acknowledgement state; the owner resumes automatic incremental review and landing.
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.

