review-repairbay
Skill1 starsChanged 4 months agoApache-2.0View on GitHub
Inspect or address actionable GitHub pull request review feedback. Use when the user wants to inspect unresolved review threads, fix selected feedback, or clear all actionable comments on a PR. Use `gh` and the bundled GraphQL script whenever thread-level state, resolution status, or inline review context matters.
In this repository19 other files it gives its agents.
- AGENTS.mdAGENTS.md
- agent-aeoSkill
- agent-memorySkill
- beanquerySkill
- ci-repairbaySkill
- delegated-developmentSkill
- ghSkill
- issue-blueprintSkill
- issue-workbenchSkill
- minifluxSkill
- og-imageSkill
- opSkill
- pencil-to-figma-migrationSkill
- pr-comment-sweepSkill
- pr-launchpadSkill
- repo-surveyorSkill
- review-checkpointSkill
- shipyardSkill
- skill-optimizerSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
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
- Distinctive: line 8
Use thread-aware GraphQL data when review state matters.
- Very specific: line 44
Only explicit clear-all authorizes replies and resolutions.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
Review Repairbay
Use thread-aware GraphQL data when review state matters; flat comments are not a complete thread model.
Memory
Direct entry loads memory only when feedback depends on prior project decisions.
Evidence-complete routine fixes skip it; inspect-only and nested invocation
defer it to the caller. Capture durable review decisions, repository rules, or
reusable root causes—not comments, thread state, IDs, or routine fixes.
Workflow
- Select intent:
inspect-onlyreads and reports;fix-selectedchanges only supplied feedback and asks for selection when absent;clear-allauthorizes scoped local fixes plus required thread replies and resolutions. - Feedback is evidence-complete when it names the file or symbol, defect, and
expected behavior. Then
fix-selectedskips PR/thread discovery: inspect the named code, direct callers, and regression tests; make the smallest fix and run focused validation. Direct mode then runs any required broader validation; nested Shipyard mode leavesfinal_checkto Shipyard. Escalate only when evidence is contradicted or insufficient. Direct fixes remain uncommitted and unpushed unless the user requested publication. When nested Shipyard explicitly delegates commit-and-push authority, inspect the final scoped code diff, commit only those fixes, and push before returning; nesting alone is not authority.fix-selectednever authorizes GitHub thread writes. Otherwise derive repo and PR from a URL, require
--repowith a number, or resolve the current branch. When discovery, inline context, replies, or resolution matter, run:python "<skill>/scripts/fetch_comments.py" --repo OWNER/REPO --pr NUMBER- Group by file or behavior; exclude informational, approved, resolved, outdated, and duplicate threads. Keep changes traceable and ask only about ambiguity, conflicts, or material product decisions.
- Only explicit
clear-allauthorizes replies and resolutions. Reply/resolve each addressed actionable thread and re-fetch until unresolved actionable count is zero. No other mode or nested caller implies thread-write authority; submitting a review or unrelated external write still requires explicit authorization.
Finish with:
status=PASS|BLOCKED|PENDING artifacts=<path-or-none> summary=<one line>PASS means the selected mode is complete and validated; clear-all also
requires the zero-thread re-fetch. BLOCKED means ambiguity, conflict, product
decision, missing scope, or unavailable GitHub access. PENDING means a
requested write or re-fetch is active. Do not print raw diffs or repeated
progress; on auth or rate limits, request reauthentication and retry.
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.

