swe-pr-reviewer

Reviews existing MRs/PRs, provides code feedback, responds to comments. Use ONLY for existing MRs - NOT for creating new ones (use swe-reviewer for that).

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "swe-pr-reviewer" with this command: npx skills add diegocanepa/agent-skills/diegocanepa-agent-skills-swe-pr-reviewer

🔍 SWE PR Reviewer

Use this skill to review existing MRs/PRs or respond to review feedback.

Do NOT use for:


Quick Start

  1. Fetch MR details - Get PR/MR info from GitHub/GitLab
  2. Analyze changes - Review files for quality + bugs + security
  3. Draft comments - For each issue: specific, actionable, reasoned
  4. MITM - Get USER approval before posting
  5. Post review - Submit with status: Approve / Changes Requested / Comment
  6. Respond to feedback (if requested) - Address comments, push changes, re-review

Analysis Checklist (Per File)

  • Architectural - Aligns with design? Separation of concerns?
  • Code Quality - SOLID? Error handling? Type safety? No code smells?
  • Bugs - Edge cases? Null checks? Race conditions? Leaks?
  • Security - Input validation? SQL injection? XSS? Secrets exposed?
  • Performance - Efficient algorithms? Unnecessary loops? Indexing?
  • Tests - Comprehensive? Edge cases? Maintainable?
  • Docs - Complex logic documented?

Comment Guidelines

Good comment:

  • Specific: Points to exact code/line
  • Actionable: Clear what needs to change
  • Reasoned: Explains WHY
  • Solution-oriented: Offers alternative or fix explanation
  • Professional: Respectful tone

Bad: "This is bad" ❌
Good: "This validation logic would benefit from extraction (lines 45-67) to make it independently testable. See: [example]" ✅


Review Statuses

Approve: Code is good, ready to merge

Changes Requested: Critical issues must be fixed first
→ Author makes changes + pushes new commits
→ Return for re-review

Comment: Feedback/questions, but not blocking


Responding to Feedback

When author makes changes:

  1. Fetch updated MR
  2. Review only the NEW changes
  3. Verify issues were addressed
  4. Re-submit review

Tools


Use Case 1: Reviewing an Existing MR

Goal: Provide comprehensive, constructive code review.

Steps:

  1. Fetch MR details

    • Use GitHub/GitLab MCP tools to get MR information:
      • MR title and description
      • Changed files
      • Diff/changes
      • Existing comments
      • Status (open, approved, changes requested)
  2. Read related context

    • Read the linked issue
    • Review architectural analysis (if available)

Use Case: Review Existing MR

  1. Fetch MR - Get details from GitHub/GitLab MCP
  2. Analyze - Check each changed file against analysis checklist
  3. Draft comments - For each issue, be specific + actionable + reasoned
  4. MITM - Show USER the comments, get approval
  5. Post - Reply to all, submit with status (Approve / Changes Requested / Comment)

Use Case: Respond to Review Feedback

  1. Fetch comments - Get all review threads
  2. Draft responses - For each: acknowledge, clarify/fix, confirm
  3. MITM - Show USER the responses, get approval
  4. Post responses - Reply to each thread
  5. Make changes (if needed) - Call swe-implementer, make fixes, push, update MR

Review Statuses

  • Approve: Code is good, ready to merge
  • Changes Requested: Blocking issues must be fixed first
  • Comment: Feedback only, not blocking
    • auth.controller.ts: 💡 Suggestion - extract validation
    • auth.test.ts: ❓ Question - missing edge case tests?
  1. Draft 4 comments (1 blocking, 1 suggestion, 1 question, 1 praise)
  2. Present to USER → Approved
  3. Post review comments
  4. Submit review: "Request Changes" due to blocking issue
  5. Author responds and fixes issue
  6. Re-review → Fixed ✅
  7. Approve MR

Example Response Flow

Responding to review on MR #456

  1. Fetch review comments - 3 comments received
  2. Analyze:
    • Comment 1: Blocking - missing salt in hashing → Need to fix
    • Comment 2: Suggestion - extract validation → Good idea
    • Comment 3: Question - edge case tests → Need to clarify
  3. Draft responses:
    • Comment 1: "Good catch! Adding salt. Will use bcrypt.genSalt(10)."
    • Comment 2: "Great suggestion! I'll extract to UserValidator."
    • Comment 3: "Added in lines 45-60, testing null email and weak passwords."
  4. Present to USER → Approved
  5. Post responses
  6. Make code changes:
    • Fix password hashing with salt
    • Extract validation logic
  7. Commit and push
  8. Update responses with commit hash
  9. Mark conversations as resolved

Specialized Skills Used

  • backend-coder - Code quality standards for review
  • vcs-issue-management - Reading linked issues
  • GitHub/GitLab MCP - Fetching MR details, posting comments, submitting reviews

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Coding

backend-coder

No summary provided by upstream source.

Repository SourceNeeds Review
Automation

skill-improvement

No summary provided by upstream source.

Repository SourceNeeds Review
Automation

mermaid-diagrams

No summary provided by upstream source.

Repository SourceNeeds Review
Automation

add-skill

No summary provided by upstream source.

Repository SourceNeeds Review