review-pr
Skill8 starsChanged 35 days agoMITView on GitHub
Review changed files or an open PR with code-reviewer, smell-reviewer, security-reviewer, and performance-reviewer. Use when the user wants a quality check on a diff, a PR number, or the current working-tree changes without running the full implement pipeline. Trigger this when someone says: review my PR, check my code, look at this diff, quality check, code review, does this look good, review PR #123, give me feedback on these changes, smell check, structural review. Do NOT use when the user wants to implement new changes — use /implement instead.
185 other projects use a version of this.
Adopted by
Most recently taken up by
- ayoubben18/ab-methodtheir own version, first seen Nov 15, 2026
- jameswyse/skillstheir own version, first seen Nov 15, 2026
- connext-solutions/flowctltheir own version, first seen Nov 15, 2026
- atomicinnovation/acceleratortheir own version, first seen Nov 14, 2026
- suitedaces/dorabottheir own version, first seen Nov 14, 2026
- AMD-AGI/Hyperloomtheir own version, first seen Nov 12, 2026
Used with
Projects with it also use diagnose-ci-failures, docs-loop 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.
zerx-labSkilldiagnose-ci-failures
Diagnose CI failures for a PR using the GitHub CLI, extract error logs, and generate a plan to fix them. Use when the user asks to check CI status, pull CI issues, triage test failures, or investigate PR build failures.
Also in 3 projects that use this
GeiserXSkilldocs-loop
Use when the user asks to audit docs, check documentation for stale claims, or verify docs against the code.
Also in 3 projects that use this
GeiserXSkillfull-loop
Drive a repository to done with sergio-loop, then keep improving it with refine-loop at a descending theta. Use only when the user explicitly invokes full-loop.
Also in 3 projects that use this
n24q02mSkillreview-delta
Review only changes since last commit using impact analysis. Token-efficient delta review with automatic blast-radius detection.
Also in 3 projects that use this
n24q02mSkillbuild-graph
Build or update the code review knowledge graph. Run this first to initialize, or let hooks keep it updated automatically.
Also in 3 projects that use this
- +7 more from GeiserX
Other versions185 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 185
- denoland/deno★ 109ksame name
- ToolJet/ToolJet★ 41ksame name
- remix-run/remix★ 33ksame name
- tirth8205/code-review-graph★ 33ksame name
- alpinejs/alpine★ 32ksame name
- nrwl/nx★ 29ksame name
- PrefectHQ/fastmcp★ 29ksame name
- jlowin/fastmcp★ 29ksame name
- livewire/livewire★ 24ksame name
- microsoft/fluentui★ 20ksame name
175 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 repository30 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- backlogSkill
- conventionsSkill
- debugSkill
- devops-azureSkill
- devops-githubSkill
- devopsSkill
- hotfixSkill
- implementSkill
- interview-meSkill
- lint-agentsSkill
- memory-auditSkill
- memory-querySkill
- obsidian-briefSkill
- obsidian-captureSkill
- obsidian-dailySkill
- obsidian-recapSkill
- obsidian-searchSkill
- obsidianSkill
- onboardSkill
- pack-reviewSkill
- planSkill
- refactorSkill
- repo-mapSkill
- scaffoldSkill
- setup-projectSkill
- smellSkill
- spec-intakeSkill
- system-checkSkill
- whatSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
14 rules. 0 are common, 11 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 178 edited versions we read keep these.
11
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 8
Run a focused multi-reviewer pass on a set of changed files or an open pull request.
- Very specific: line 14
GitHub PR: Before fetching anything, apply devops-github's setup verification and repo-targeting steps (skills/devops-github/SKILL.md §1–2) …
- Distinctive: line 16
If the user supplied file paths, use those directly.
- Distinctive: line 17
If neither was supplied, run git diff main...HEAD --name-only (or git diff HEAD --name-only if on main) to discover the changed files.
- Very specific: line 46
Do not run git merge/git merge --abort to test this — it mutates the working tree and can strand the user mid-merge.
- Distinctive: line 50
Before dispatching reviewers, classify what the changes touch:
- Distinctive: line 65
Run code-reviewer, smell-reviewer, security-reviewer, and performance-reviewer in parallel — their inputs are independent.
- Distinctive: line 69
After all reviewers complete, produce a consolidated report:
- Distinctive: line 106
If there are Critical findings: recommend routing back to the responsible engineer before merging.
- Distinctive: line 107
If there are only Minor/Advisory findings: surface them and ask the user whether to fix or accept.
- Distinctive: line 108
If no findings: confirm the PR looks clean and suggest merging when ready.
- Distinctive: line 109
If the mergeability probe reported conflicts: call this out prominently even when the code review is clean — a clean review does not mean …
- Very specific: line 115
Reviewing stale diffs: If the PR has been updated since the user last looked at it, re-fetch the diff (gh pr diff <number> or the git diff …
- Distinctive: line 117
Auth/setup failures, ambiguous repo or project targeting: Don't troubleshoot these here — they're devops-github's and devops-azure's …
“Common” is measured: kept by at least 6 in 10 of the 178 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 PR / Changed Files
Run a focused multi-reviewer pass on a set of changed files or an open pull request.
1. Determine the change set
If the user supplied a PR number, first resolve which platform it's on, using the same signals as
/devops: repo name matchingGITHUB_REPOS, project name matchingAZURE_DEVOPS_PROJECTS, "work item"/"board" language, or the local git remote's host. If signals conflict or neither is present, ask directly rather than guessing.- GitHub PR: Before fetching anything, apply devops-github's setup verification and repo-targeting steps (
skills/devops-github/SKILL.md§1–2) — confirmghis installed and authenticated, and resolve the target repo fromGITHUB_REPOS/local git remote. Then rungh pr diff <number> --repo <org>/<repo>for the diff andgh pr view <number> --repo <org>/<repo>for the title, description, and author. - Azure DevOps PR: Apply devops-azure's setup verification and org/project-resolution steps (
skills/devops-azure/SKILL.md§1, §3–4) — confirmazand thedevopsextension are installed and authenticated, and resolve the org (PR IDs are org-scoped, so per devops-azure §2 only the org is strictly required). Fetch metadata withaz repos pr show --id <number> --org https://dev.azure.com/<org>. TheazCLI has no direct diff command — read the source/target branches from that output,git fetchboth, and rungit diff <target>...<source>locally to get the diff.
- GitHub PR: Before fetching anything, apply devops-github's setup verification and repo-targeting steps (
- If the user supplied file paths, use those directly.
- If neither was supplied, run
git diff main...HEAD --name-only(orgit diff HEAD --name-onlyif on main) to discover the changed files. Confirm the list with the user before proceeding.
1a. Probe mergeability against the target branch
Advisory check — surfaces merge conflicts before the PR reaches GitHub/Azure's automatic review. Run it whenever there is a branch to compare (a PR, or a working-tree review off a feature branch). Skip it for loose-file-path reviews with no branch context.
Resolve the two refs:
- GitHub PR: target and head come from
gh pr view <number> --repo <org>/<repo> --json baseRefName,headRefName. Fetch both first — they may not exist locally:git fetch <remote> <baseRefName> <headRefName>. - Azure DevOps PR: you already read the source (head) and target branches from
az repos pr showin Step 1, and alreadygit fetched them to reconstruct the diff. Reuse those refs. - Working-tree / branch review: head is the current branch; the target is the repo default — prefer
origin/HEAD, falling back to a localmainthenmaster:base=$(git symbolic-ref -q refs/remotes/origin/HEAD 2>/dev/null | sed 's#^refs/remotes/##') [ -z "$base" ] && for c in main master; do git rev-parse --verify -q "$c" >/dev/null && { base="$c"; break; }; done
Probe without touching the working tree or index using git merge-tree (Git 2.38+):
git merge-tree --write-tree --name-only "<target>" "<head>"; echo "exit=$?"- exit 0: merges cleanly — report Mergeable.
- exit 1: conflicts — stdout is the synthetic tree OID on line 1, then the conflicting paths. Report those paths.
- exit 128 / unknown option: older Git without
--write-tree. Fall back to the legacy three-arg form, which always exits 0 but prints conflict hunks to stdout:git merge-tree "$(git merge-base <target> <head>)" "<target>" "<head>" | grep -q '^<<<<<<<' && echo "CONFLICTS" || echo "CLEAN"
Do not run git merge/git merge --abort to test this — it mutates the working tree and can strand the user mid-merge. merge-tree is read-only by design.
2. Classify the change set
Before dispatching reviewers, classify what the changes touch:
- Always invoke: code-reviewer, smell-reviewer (for any code changes — skip for docs/config/migrations only)
- Invoke if changes touch authentication, authorization, data access, PII, external endpoints, or secrets: security-reviewer
- Invoke if changes include database queries, API endpoints, loops over collections, or caching logic: performance-reviewer
State which reviewers you are invoking and why before starting.
3. Dispatch reviewers (in parallel)
Pass each reviewer:
- The list of changed file paths
- The PR title and description (if available)
- Any relevant context the user provided
Run code-reviewer, smell-reviewer, security-reviewer, and performance-reviewer in parallel — their inputs are independent.
4. Synthesize findings
After all reviewers complete, produce a consolidated report:
## Review Summary — <PR title or branch name>
### Mergeability
- <Mergeable into `<target>`, no conflicts.> — OR —
- <Conflicts with `<target>` in: path/a, path/b. Resolve before merging.> — OR —
- <Not checked: loose-file review with no branch context.>
### Critical
- <finding> [source: code-reviewer | security-reviewer | performance-reviewer | smell-reviewer]
### Major
- <finding> [source: ...]
### Minor / Advisory
- <finding> [source: ...]
### No findings
- <reviewer> found no issues.Group by severity (Critical → Major → Minor). Include the source reviewer for each finding.
Map each reviewer's severity vocabulary into the three buckets:
| Bucket | code-reviewer / smell-reviewer | security-reviewer | performance-reviewer |
| Critical | Critical | Critical, High | — |
| Major | Warning | Medium | High |
| Minor / Advisory | Suggestion | Low | Medium, Low |
(security High maps to Critical because it blocks merge-reviewer; performance findings are always advisory, so even High is Major at most.)
5. Recommend next steps
- If there are Critical findings: recommend routing back to the responsible engineer before merging.
- If there are only Minor/Advisory findings: surface them and ask the user whether to fix or accept.
- If no findings: confirm the PR looks clean and suggest merging when ready.
- If the mergeability probe reported conflicts: call this out prominently even when the code review is clean — a clean review does not mean the branch will merge. Recommend rebasing onto / merging the target branch and resolving conflicts before opening or updating the PR.
Gotchas
- No PR number and no files supplied: Running
git diff main...HEAD --name-onlyon a fresh branch returns nothing. Confirm with the user that the branch has commits before proceeding. - Skipping security-reviewer for API changes: Even small endpoint changes can introduce authorization gaps. When in doubt, include security-reviewer — it is faster to run it than to explain why you didn't.
- Reviewing stale diffs: If the PR has been updated since the user last looked at it, re-fetch the diff (
gh pr diff <number>or thegit diffreconstructed from Azure DevOps branches) rather than relying on a diff the user pasted. Cached diffs miss new commits. - Conflating review with implementation: This skill surfaces findings — it does not fix them. If the user asks you to fix a Critical finding during review, pause and confirm whether they want to switch to /implement or /debug.
- Auth/setup failures, ambiguous repo or project targeting: Don't troubleshoot these here — they're devops-github's and devops-azure's concern. Follow their gotchas sections (
skills/devops-github/SKILL.md,skills/devops-azure/SKILL.md) instead of improvising a workaround. - Azure DevOps diff reconstruction fails (branch not found locally):
git fetchthe source/target refs from the PR's remote (az repos pr showoutput includes the repository) before diffing — don't assume they're already present locally. - Mergeability probe on unfetched refs:
git merge-treecompares whatever the refs point at locally. If you probe a stalemain(or a head ref you never fetched), the result is wrong — a phantom conflict or a false "clean". Alwaysgit fetchboth refs immediately before the probe, and probe the remote-tracking ref (origin/main), not a stale localmain. - Reporting mergeability as a code finding: Conflicts against the target are a state of the branch, not a code-quality defect — keep them in the dedicated Mergeability section, not mixed into Critical/Major. A branch can be perfectly written and still conflict because the target moved.
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.

