review-panel
Skill0 starsChanged 16 days agoMITView on GitHub
Use when several reviewers should check the same change, spec, or document independently, or when the same review must sometimes only report (validate) and sometimes also apply fixes (finalize). Routes to code-council (code mode for code, plan mode for specs and design docs) with your lens list, so the panel gets its refuters, Opus chairman and failure handling.
13 other projects use a version of this.
Adopted by
Most recently taken up by
- JeffreyFerreiras/ai-skillstheir own version, first seen Nov 1, 2026
- rosslevinsky/portable-agent-skillstheir own version, first seen Oct 15, 2026
- gigio1023/agent-skillstheir own version, first seen Oct 13, 2026
- gigio1023/gigio-skillstheir own version, first seen Oct 13, 2026
- raine/consult-llm-mcptheir own version, first seen Oct 8, 2026
- HACK-WU/skillstheir own version, first seen Sep 25, 2026
Other versions13 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 13
- raine/consult-llm-mcp★ 148same name
- 9thLevelSoftware/legion★ 76same name
- iamneilroberts/claude-skills★ 26same name
- rosslevinsky/portable-agent-skills★ 12same name
- josherau/claude-operating-core★ 3same name
- gigio1023/agent-skills★ 2same name
- gigio1023/gigio-skills★ 2same name
- HACK-WU/skills★ 2same name
- kbichave/skills★ 2same name
- JeffreyFerreiras/ai-skills★ 1same name
3 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 repository23 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- agent-orchestrationSkill
- ask-for-approvalSkill
- claim-verification-loopSkill
- code-councilSkill
- code-reviewSkill
- context-as-imagesSkill
- grillingSkill
- grow-test-coverageSkill
- investigateSkill
- loop-until-verifiedSkill
- plan-and-buildSkill
- prompt-cachingSkill
- proofreadSkill
- review-allSkill
- review-before-mergeSkill
- to-questionnaireSkill
- track-review-findingsSkill
- trim-contextSkill
- unslopSkill
- verify-findingsSkill
- whatnowSkill
- wizardSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 12 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 24
Lint first: Run a cheap deterministic check (lint, types, schema, missing fields, duplicate IDs, placeholders).
- Distinctive: line 27
Pick lenses: Always one general lens against the artifact's own acceptance criteria, plus lenses the risk calls for (security, data …
“Common” is measured: kept by at least 6 in 10 of the 12 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
Review panel
Usually called through
/proofread. Follow~/.claude/skills/proofread/proof-rules.md(proof, verdict words, Sonnet workers, Opus adjudicators, report location). Where this skill and those rules differ, the stricter one wins.
Overview
This skill is a thin entry point. It does not run its own reviewers. It runs code-council
in code mode, or plan mode for a spec or design doc (~/.claude/skills/code-council/SKILL.md), with your lenses, so every finding
goes through code-council's batched refuters, its Opus chairman, and its rule that a dead
agent is re-run once and never read as "no findings".
Skip it for a one-line change, a throwaway prototype, or when a deterministic check already answers the question.
The pattern
- Lint first. Run a cheap deterministic check (lint, types, schema, missing fields, duplicate IDs, placeholders). It catches objective errors free, before any agent spends judgment on them.
- Pick lenses. Always one general lens against the artifact's own acceptance criteria,
plus lenses the risk calls for (security, data integrity, cross-team). Light for a small
change, full for a high-risk one. Code-council takes them as its optional
lenseslist ([{ key, label, focus }]) and replaces its default four. - Run code-council in code mode (plan mode for a spec or design doc) with that list and the target. Workers stay on Sonnet, the chairman on Opus, at most 5 agents at once.
- Branch on mode.
- Validate: report only. Never change the artifact.
- Finalize: apply only the fixes a refuter confirmed. Then re-run the lint and the project's checks and read the result back to confirm each fix took effect. Then run code-council once more on the fix diff, as a fresh review, so the author does not grade its own fix. At most 3 fix rounds, then list what is left. Anything REFUTED, UNVERIFIED or needing a human call is listed, not applied.
Stop condition: code-council returned its report, every lens has a verdict or is listed as not run, and in finalize mode the re-run checks passed.
Common mistakes
- Running reviewers sequentially; a reviewer who sees an earlier opinion is no longer independent.
- Treating an errored or timed-out reviewer as if it found nothing.
- Applying a fix nobody tried to disprove.
- Calling a fix done without re-running the checks.
- Running the full panel on every small change instead of scaling to stakes.
Source
Adapted from skills/bmad-architecture/references/reviewer-gate.md in
https://github.com/bmad-code-org/BMAD-METHOD (commit 1b59caa7f96459fda6750c225a9330283e108fd6;
own words only, see that repo's LICENSE). Kept: check-first order and the finalize-vs-validate split.
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.