review-pull-requests — for Claude Code review-pull-requests, community, for Claude Code, ide skills, pull_request_read, method: get, closingIssuesReferences, require_conventional_commits, require_linked_issue, Closes #N

v1.0.0

Über diesen Skill

Geeigneter Einsatz: Ideal for AI agents that need review pull requests. Lokalisierte Zusammenfassung: Intelligent development portals # Review Pull Requests Reviews GitHub pull requests for completeness, quality, and merge readiness according to repository conventions. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

Funktionen

Review Pull Requests
Reviews GitHub pull requests for completeness, quality, and merge
readiness according to repository conventions.
Step 1: Fetch PR Metadata
Use GitHub MCP pull request read with method: get to retrieve:

# Core Topics

ourchitecture ourchitecture
[1]
[0]
Updated: 4/24/2026

Killer-Skills Review

Decision support comes first. Repository text comes second.

Reference-Only Page Review Score: 10/11

This page remains useful for teams, but Killer-Skills treats it as reference material instead of a primary organic landing page.

Original recommendation layer Concrete use-case guidance Explicit limitations and caution Quality floor passed for review
Review Score
10/11
Quality Score
57
Canonical Locale
en
Detected Body Locale
en

Geeigneter Einsatz: Ideal for AI agents that need review pull requests. Lokalisierte Zusammenfassung: Intelligent development portals # Review Pull Requests Reviews GitHub pull requests for completeness, quality, and merge readiness according to repository conventions. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

Warum diese Fähigkeit verwenden

Empfehlung: review-pull-requests helps agents review pull requests. Intelligent development portals # Review Pull Requests Reviews GitHub pull requests for completeness, quality, and merge readiness according to

Am besten geeignet für

Geeigneter Einsatz: Ideal for AI agents that need review pull requests.

Handlungsfähige Anwendungsfälle for review-pull-requests

Anwendungsfall: Applying Review Pull Requests
Anwendungsfall: Applying Reviews GitHub pull requests for completeness, quality, and merge
Anwendungsfall: Applying readiness according to repository conventions

! Sicherheit & Einschränkungen

  • Einschraenkung: require conventional commits is true:
  • Einschraenkung: If require linked issue is true, at least one issue reference
  • Einschraenkung: If require linked issue is true:

Why this page is reference-only

  • - Current locale does not satisfy the locale-governance contract.

Source Boundary

The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.

After The Review

Decide The Next Action Before You Keep Reading Repository Material

Killer-Skills should not stop at opening repository instructions. It should help you decide whether to install this skill, when to cross-check against trusted collections, and when to move into workflow rollout.

Labs Demo

Browser Sandbox Environment

⚡️ Ready to unleash?

Experience this Agent in a zero-setup browser environment powered by WebContainers. No installation required.

Boot Container Sandbox

FAQ & Installation Steps

These questions and steps mirror the structured data on this page for better search understanding.

? Frequently Asked Questions

What is review-pull-requests?

Geeigneter Einsatz: Ideal for AI agents that need review pull requests. Lokalisierte Zusammenfassung: Intelligent development portals # Review Pull Requests Reviews GitHub pull requests for completeness, quality, and merge readiness according to repository conventions. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

How do I install review-pull-requests?

Run the command: npx killer-skills add ourchitecture/idp/review-pull-requests. It works with Cursor, Windsurf, VS Code, Claude Code, and 19+ other IDEs.

What are the use cases for review-pull-requests?

Key use cases include: Anwendungsfall: Applying Review Pull Requests, Anwendungsfall: Applying Reviews GitHub pull requests for completeness, quality, and merge, Anwendungsfall: Applying readiness according to repository conventions.

Which IDEs are compatible with review-pull-requests?

This skill is compatible with Cursor, Windsurf, VS Code, Trae, Claude Code, OpenClaw, Aider, Codex, OpenCode, Goose, Cline, Roo Code, Kiro, Augment Code, Continue, GitHub Copilot, Sourcegraph Cody, and Amazon Q Developer. Use the Killer-Skills CLI for universal one-command installation.

Are there any limitations for review-pull-requests?

Einschraenkung: require conventional commits is true:. Einschraenkung: If require linked issue is true, at least one issue reference. Einschraenkung: If require linked issue is true:.

How To Install

  1. 1. Open your terminal

    Open the terminal or command line in your project directory.

  2. 2. Run the install command

    Run: npx killer-skills add ourchitecture/idp/review-pull-requests. The CLI will automatically detect your IDE or AI agent and configure the skill.

  3. 3. Start using the skill

    The skill is now active. Your AI agent can use review-pull-requests immediately in the current project.

! Reference-Only Mode

This page remains useful for installation and reference, but Killer-Skills no longer treats it as a primary indexable landing page. Read the review above before relying on the upstream repository instructions.

Upstream Repository Material

The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.

Upstream Source

review-pull-requests

Install review-pull-requests, an AI agent skill for AI agent workflows and automation. Review the use cases, limitations, and setup path before rollout.

SKILL.md
Readonly
Upstream Repository Material
The section below is imported from the upstream repository and should be treated as secondary evidence. Use the Killer-Skills review above as the primary layer for fit, risk, and installation decisions.
Supporting Evidence

Review Pull Requests

Reviews GitHub pull requests for completeness, quality, and merge readiness according to repository conventions.

Step 1: Fetch PR Metadata

Use GitHub MCP pull_request_read with method: get to retrieve:

  • PR number, title, body
  • Author, assignees, reviewers
  • State (open, closed, merged), draft status
  • Base and head branches
  • Labels
  • Linked issues (from closingIssuesReferences or parsed from body)
  • Created and updated timestamps

Fallback command if MCP is unavailable:

bash
1gh pr view <pr_number> --json number,title,body,state,isDraft,headRefName,baseRefName,author,assignees,labels,commits,reviews,additions,deletions,changedFiles,url,createdAt,updatedAt

If the PR is not found, stop and report an error.

Step 2: Validate PR Metadata

Check PR title and body:

  1. Title: Verify it follows Conventional Commits format if require_conventional_commits is true:

    • Pattern: ^(feat|fix|docs|style|refactor|perf|test|build|ci|chore|security|revert)(\(.+\))?!?: .+$
    • Max 72 characters
    • Lowercase first word after colon
    • No trailing period
  2. Body: Verify it contains:

    • A clear description of what changed and why.
    • If require_linked_issue is true, at least one issue reference (Closes #N, Fixes #N, Resolves #N, or Refs #N).
    • If the change is breaking, a BREAKING CHANGE: footer or ! in the title.
  3. Labels: Check that the PR has appropriate labels (e.g., bug, enhancement, documentation, chore).

  4. Draft Status: If the PR is still a draft, note this in the report. Full review continues but recommendations defer to "mark ready for review" before merging.

Step 3: Validate Linked Issues

If require_linked_issue is true:

  1. Parse issue references from the PR body and title.

  2. For each referenced issue:

    • Verify the issue exists and is open (or recently closed if the PR was the closing action).
    • Check that the issue has the ready label (not needs-triage or blocked).
    • Verify that the PR scope aligns with the issue description.
  3. If no issues are linked, add a high-severity finding.

  4. If multiple issues are linked, verify they are related and the PR scope is appropriate.

Step 4: Review Commit History

Use GitHub MCP pull_request_read with method: get to retrieve commit list, or use fallback:

bash
1gh pr view <pr_number> --json commits --jq '.commits[].messageHeadline'

Validate commits:

  1. If require_conventional_commits is true:

    • Check that each commit message follows Conventional Commits format.
    • Flag commits with unclear or generic messages ("fix", "update", "wip").
  2. Check for:

    • Merge commits (should be rare; squash merges are preferred).
    • Revert commits without explanation.
    • Commits that add and then immediately remove the same code (churn).
  3. Recommend squashing fixup commits or clarifying commit intent if the history is confusing.

Step 5: Review Changed Files

Use GitHub MCP pull_request_read with method: get_files to list all changed files, or use fallback:

bash
1gh pr diff <pr_number> --name-only

Validate changes:

  1. Source code changes:

    • If require_tests is true and source code was modified, verify that test files were also added or updated.
    • Flag large diffs (e.g., 500+ lines changed in a single file) without clear justification.
  2. Documentation changes:

    • If source code, APIs, or configuration changed, verify that corresponding documentation was updated.
    • Check that README, AGENTS.md, or docs/ files reflect the changes.
  3. Configuration or infrastructure changes:

    • If Makefile, moon.yml, .github/workflows, Dockerfile, or package.json changed, verify the changes are intentional and documented.
  4. Sensitive files:

    • Flag changes to .env files, secret managers, or credential storage.
    • Verify no secrets or credentials are committed in the diff.

Step 6: Check CI Status (conditional)

If check_ci_status is true:

  1. Use GitHub MCP pull_request_read with method: get_check_runs to list all CI checks.

  2. Verify that all required checks have passed:

    • Conclusion must be success for required checks.
    • Flag checks with failure, cancelled, timed_out, action_required, stale, or startup_failure.
  3. If any required checks are still pending or in progress, note this in the report and defer merge recommendation until checks complete.

Fallback command:

bash
1gh pr checks <pr_number>

Step 7: Review Existing PR Comments and Reviews

Use GitHub MCP pull_request_read with method: get_reviews, method: get_review_comments, and method: get_comments to fetch existing reviews and comments.

Check:

  1. Unresolved review comment threads: If review comment threads exist and are not resolved, flag this as a high-severity blocker to merge. Each unresolved thread must be addressed by either:

    • Making the requested code change and resolving the thread
    • Replying with clarification and resolving if the reviewer accepts
    • Discussing and reaching consensus before resolving
  2. Review status:

    • Count approvals (APPROVED), rejections (CHANGES_REQUESTED), and comment-only reviews (COMMENTED).
    • Verify that at least one CODEOWNERS-designated reviewer has approved if repository rulesets require it.
    • If any reviewer has requested changes, flag as high-severity until those changes are addressed and the reviewer re-reviews.
  3. Stale reviews: If reviews are older than the latest commit, suggest re-requesting review (medium severity).

  4. General PR comments: Review general issue comments on the PR (not code review comments) for guidance, questions, or blockers. These comments do not have the same authority as code review comments but may contain important context:

    • Flag unanswered questions from maintainers as medium severity.
    • Note any blockers or dependencies mentioned in comments.
    • Treat general comments as advisory unless they come from CODEOWNERS or contain explicit blocking language.

Fallback:

bash
1gh pr view <pr_number> --json reviews,comments --jq '.reviews[] | {author: .author.login, state: .state, submittedAt: .submittedAt}'

Step 7a: Check Merge Conflicts and Branch Status

Use GitHub MCP pull_request_read with method: get_status or gh CLI to check:

  1. Merge conflicts: Verify the PR branch can be cleanly merged into the base branch. If conflicts exist, flag as high-severity blocker.

    bash
    1gh pr view <pr_number> --json mergeable,mergeStateStatus
    • mergeable: CONFLICTING indicates merge conflicts
    • mergeStateStatus: DIRTY indicates conflicts or other issues
  2. Branch update needed: Check if the base branch has advanced past the PR head and the PR branch needs updates:

    • Compare base branch HEAD with the merge base of the PR
    • If the PR is behind the base branch and repository policy requires up-to-date branches, flag as medium severity
    • Recommend updating the branch via merge or rebase
  3. Branch protection compliance: Verify the PR satisfies all branch protection rules:

    • Required status checks passing
    • Required reviews obtained
    • No unresolved conversations (if required by ruleset)
    • Linear history requirement (squash merge ready)

Fallback:

bash
1gh pr view <pr_number> --json mergeable,mergeStateStatus,statusCheckRollup

Step 8: Score and Classify Findings

Assign severity to each finding:

  • High: Missing linked issue (when required), CI checks failing, secrets in diff, unresolved review threads, merge conflicts, changes requested by reviewers, breaking changes without documentation.
  • Medium: Non-conventional commit messages, missing tests (when required), missing documentation updates, large diffs without justification, draft PR not marked ready, branch needs update, unanswered maintainer questions in comments, stale reviews.
  • Low: Minor title/body formatting issues, missing labels, advisory comments from non-reviewers.

Calculate a quality score starting from 100:

  • High: -25 each (max penalty 75)
  • Medium: -12 each (max penalty 36)
  • Low: -5 each (max penalty 20)

Clamp to [0, 100].

Set overall_status:

  • fail if any high-severity finding exists
  • warn if no high findings but medium/low findings exist
  • pass if no findings exist

Set merge_ready:

  • true only if:
    • overall_status is pass
    • All required CI checks have passed
    • PR is not a draft
    • At least one approving review exists (if required by rulesets)
    • No unresolved review conversations
    • No merge conflicts
    • No reviewers have requested changes
    • Branch is up to date with base (if required by repository policy)

Step 9: Generate Review Report

Produce a structured report:

  1. Summary:

    • PR number, title, author, branch
    • State, draft status, linked issues
    • CI check status
    • Merge status (conflicts, branch up to date)
    • Review status (approvals, changes requested, unresolved threads)
    • Comment resolution status (resolved vs unresolved review comments)
    • Overall status and score
  2. Findings: Each finding with severity, category, description, and recommended action.

  3. Merge Recommendation:

    • If merge_ready is true: "PR is ready to merge."
    • Otherwise: List blockers and recommended next steps.

Example finding structure:

json
1{ 2 "severity": "high", 3 "category": "linked-issue", 4 "message": "PR does not reference any GitHub issue", 5 "suggestion": "Add 'Closes #N' or 'Refs #N' to the PR body" 6}

Step 10: Post PR Review Comment (conditional)

If post_review is true:

  1. Format the review report as GitHub-flavored Markdown.
  2. Use GitHub MCP pull_request_review_write with method: create and event: COMMENT to post the review.

Fallback:

bash
1gh pr comment <pr_number> --body "<review_report_markdown>"

The comment should start with:

markdown
1## 🤖 Pull Request Review 2 3> Posted by AI agent. Review findings below.

If merge_ready is true and no high/medium findings exist, post an approving review with event: APPROVE. Otherwise, post as a comment only.

Step 11: Return Outputs

Return:

  • findings array
  • overall_status
  • merge_ready
  • score

If the PR passes all checks, return merge_ready: true with a score of 100 and an explicit passing statement.

Definition of Done

This skill is complete when:

  1. The PR has been fully reviewed across metadata, commits, changes, tests, docs, and CI status.
  2. Findings are categorized by severity with actionable guidance.
  3. A merge-readiness determination has been made.
  4. If post_review was true, the review was successfully posted to the PR.

Verwandte Fähigkeiten

Looking for an alternative to review-pull-requests or another community skill for your workflow? Explore these related open-source skills.

Alle anzeigen

openclaw-release-maintainer

Logo of openclaw
openclaw

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

333.8k
0
Künstliche Intelligenz

widget-generator

Logo of f
f

Erzeugen Sie anpassbare Widget-Plugins für das Prompts.Chat-Feed-System

149.6k
0
Künstliche Intelligenz

flags

Logo of vercel
vercel

Das React-Framework

138.4k
0
Browser

pr-review

Logo of pytorch
pytorch

Tensor und dynamische neuronale Netze in Python mit starker GPU-Beschleunigung

98.6k
0
Entwickler