Claude Code Plugins

Community-maintained marketplace

Feedback

Reviews pull requests created by GitHub Copilot agents before merging. Triggers on "review this PR", "check PR #123", "validate agent work", or automatically when agents complete work. Ensures quality gates are met.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name copilot-pr-reviewer
description Reviews pull requests created by GitHub Copilot agents before merging. Triggers on "review this PR", "check PR
allowed-tools Read, Grep, Glob, Bash, mcp__github__*

GitHub Copilot PR Reviewer

This skill reviews pull requests created by GitHub Copilot agents to ensure quality, correctness, and alignment with project standards before merging.

When to Use This Skill

Auto-trigger when:

  • User says: "Review PR #123"
  • User says: "Check this pull request"
  • User says: "Validate agent work"
  • User says: "Is this PR ready to merge?"
  • After a Copilot agent completes work and opens a PR
  • Before merging any agent-created PR

Manual trigger:

  • Final review before merging
  • When PR has been updated after review comments
  • Periodic review of open PRs

Mission

Ensure all agent-created pull requests meet quality standards:

  1. Code correctness - Does it do what it should?
  2. Test coverage - Are changes tested?
  3. No regressions - Did it break anything?
  4. Style compliance - Follows project conventions?
  5. Security - No vulnerabilities introduced?
  6. Performance - No performance degradation?

Review Checklist

See checklists/ directory for domain-specific checklists:

  • frontend-review.md - React, UI, components
  • backend-review.md - API, database, server
  • testing-review.md - Test quality and coverage
  • security-review.md - Security considerations
  • performance-review.md - Performance impact

Review Process (Using GitHub MCP)

1. Fetch PR & Changes

Use GitHub MCP to get PR details:

// Get PR metadata
const pr = await mcp__github__pull_request_read({
  method: "get",
  owner: "{owner}",
  repo: "{repo}",
  pullNumber: prNumber
})

// Get changed files
const files = await mcp__github__pull_request_read({
  method: "get_files",
  owner: "{owner}",
  repo: "{repo}",
  pullNumber: prNumber
})

// Get CI/CD status
const checks = await mcp__github__pull_request_read({
  method: "get_status",
  owner: "{owner}",
  repo: "{repo}",
  pullNumber: prNumber
})

Verify PR has:

  • Clear title describing what was changed
  • Description explaining why
  • Links to related issue(s)
  • Appropriate labels
  • No merge conflicts (pr.mergeable === true)
  • CI/CD checks passing (checks.every(c => c.conclusion === "success"))

2. Code Review with Inline Comments

Create pending review:

await mcp__github__pull_request_review_write({
  method: "create",
  owner: "{owner}",
  repo: "{repo}",
  pullNumber: prNumber,
  body: "Reviewing changes..."
})

Analyze each file using domain-specific checklist:

  • Changes match issue requirements
  • Code is readable and maintainable
  • No obvious bugs or issues
  • Error handling is appropriate
  • TypeScript types are correct
  • No console.logs or debug code left in

Add inline comments for issues:

await mcp__github__add_comment_to_pending_review({
  owner: "{owner}",
  repo: "{repo}",
  pullNumber: prNumber,
  path: "client/src/components/Profile.tsx",
  body: "🚫 **Blocking**: Remove console.log before merging.",
  line: 42,
  side: "RIGHT",
  subjectType: "LINE"
})

3. Testing Review

Verify:

  • Tests added for new functionality
  • Tests updated for changed functionality
  • All tests passing
  • Edge cases covered
  • Test coverage adequate (no major gaps)

4. Impact Analysis

Check:

  • No breaking changes (or documented if intentional)
  • Database migrations work correctly
  • API contracts maintained
  • Dependencies updated safely
  • Performance impact acceptable

5. Security Review

For changes involving:

  • Authentication/authorization logic
  • User input handling
  • Database queries
  • External API calls
  • File uploads
  • Sensitive data

Verify no new vulnerabilities introduced.

6. Final Checks

Before approval:

  • npm run check passes (TypeScript)
  • No new ESLint warnings
  • Documentation updated if needed
  • CLAUDE.md updated if new patterns introduced

Review Outcomes (Submit via GitHub MCP)

✅ Approve

When: PR meets all quality standards, inline comments added

Submit approval:

await mcp__github__pull_request_review_write({
  method: "submit_pending",
  owner: "{owner}",
  repo: "{repo}",
  pullNumber: prNumber,
  body: `## Review: APPROVED ✅

This PR looks great! All quality checks passed.

**Verified**:
- ✅ Code correctness
- ✅ Test coverage
- ✅ No regressions
- ✅ Style compliance
- ✅ Security
- ✅ Performance

**Ready to merge**: Yes
**Suggested next steps**: Merge and close #${issueNumber}`,
  event: "APPROVE"
})

🔄 Request Changes

When: PR has blocking issues (already added as inline comments)

Submit change request:

await mcp__github__pull_request_review_write({
  method: "submit_pending",
  owner: "{owner}",
  repo: "{repo}",
  pullNumber: prNumber,
  body: `## Review: CHANGES REQUESTED 🔄

Found ${blockingIssues.length} blocking issue(s) that need attention.

All issues have been marked inline with specific line comments.

### Summary
- 🚫 Blocking: ${blockingIssues.length}
- 💡 Suggestions: ${suggestions.length}

Please address the blocking issues and update the PR.`,
  event: "REQUEST_CHANGES"
})

Note: Individual issues are already added as inline comments via add_comment_to_pending_review

⚠️ Needs Discussion

PR has architectural or approach concerns:

## Review: DISCUSSION NEEDED ⚠️

The implementation works, but I have concerns about the approach.

### Concerns

1. **[Concern 1]**
   - What: [What's concerning]
   - Why: [Why it's a problem]
   - Alternatives: [Alternative approaches]

2. **[Concern 2]**
   - What: [What's concerning]
   - Why: [Why it's a problem]
   - Alternatives: [Alternative approaches]

**Recommendation**: Let's discuss the approach before proceeding.

@[User] please weigh in on preferred approach.

Agent Feedback

Provide constructive feedback to improve agent work:

Positive Feedback

**Nice work on**:
- Clean separation of concerns in the API layer
- Comprehensive test coverage
- Good error handling patterns

Improvement Areas

**For future PRs**:
- Consider extracting this 50-line function into smaller units
- Add JSDoc comments for complex logic
- Use `const` instead of `let` where possible

Integration with Workflow

In Multi-Agent Workflows

Review happens at transition points:

Phase 1: Backend
  ↓
  PR #200 created
  ↓
  🔍 PR Review (this skill)
  ↓
  ✅ Approved & Merged
  ↓
Phase 2: Frontend (unblocked)

Review Triggers

  • After agent completes work - Review before next phase
  • Before workflow continues - Validate checkpoint
  • On PR update - Re-review after changes
  • Manual request - User asks for review

Quality Gates

Must Pass to Merge

  1. All tests passing - CI/CD green
  2. No TypeScript errors - npm run check clean
  3. No blocking issues - All critical items resolved
  4. Security reviewed - If touching auth/data/APIs
  5. Performance acceptable - No major degradation

Should Pass to Merge

  1. Code coverage maintained - No significant drops
  2. Documentation updated - If needed
  3. Mobile responsive - For UI changes
  4. Accessibility maintained - For UI changes

Special Review Cases

Database Migrations

Extra checks:

  • Migration is reversible
  • No data loss
  • Tested on copy of prod data
  • Indexes added where needed
  • Migration runs quickly (<1 min)

API Changes

Extra checks:

  • Backward compatible or version bumped
  • Documentation updated
  • Error responses documented
  • Rate limiting considered

Security-Related Changes

Extra checks:

  • Security Specialist reviewed
  • No hardcoded secrets
  • Input validated
  • Authorization checked
  • Sensitive data encrypted

Performance-Critical Changes

Extra checks:

  • Performance Specialist reviewed
  • Benchmarks run
  • No N+1 queries introduced
  • Caching strategy sound

Reference Documentation

  • Checklists: See checklists/ directory
  • Review templates: See reference/review-templates.md
  • Common issues: See reference/common-issues.md

What This Skill Does

  • Reviews PRs for quality and correctness
  • Checks tests and coverage
  • Validates security and performance
  • Provides constructive feedback
  • Approves or requests changes
  • Gates workflow progression

What This Skill Doesn't Do

  • Doesn't write code (agents do that)
  • Doesn't fix issues (requests agent to fix)
  • Doesn't make architectural decisions (discusses with user)
  • Doesn't merge PRs (recommends merge, user decides)

Output Format

Review Summary

# PR Review Summary: #[PR-NUM]

**PR Title**: [Title]
**Issue**: #[issue-num]
**Agent**: @[Agent-Name]-Specialist
**Changes**: [Brief description]

## Review Status: [APPROVED ✅ | CHANGES REQUESTED 🔄 | DISCUSSION NEEDED ⚠️]

### Code Quality: [✅ | ⚠️ | ❌]
[Comments]

### Test Coverage: [✅ | ⚠️ | ❌]
[Comments]

### Security: [✅ | ⚠️ | ❌]
[Comments]

### Performance: [✅ | ⚠️ | ❌]
[Comments]

## Blocking Issues
[List or "None"]

## Suggestions
[List or "None"]

## Next Steps
[What should happen next]

**Reviewed by**: copilot-pr-reviewer skill
**Reviewed at**: [Timestamp]