Add GitLab MR Reviewer skill for automated code review
Summary
This MR adds a new GitLab MR Reviewer skill that enables AI-assisted code review for GitLab merge requests. The skill follows GitLab's official code review guidelines and implements conventional comments format for structured, actionable feedback.
What Changed
New Skill Added: gitlab-mr-reviewer
Main Skill File:
-
skills/gitlab-mr-reviewer/SKILL.md- Complete skill definition with workflows and examples
Reference Documentation:
-
skills/gitlab-mr-reviewer/references/api_reference.md- GitLab API tools for MR reviews -
skills/gitlab-mr-reviewer/references/code_review_guidelines.md- GitLab review standards and checklists -
skills/gitlab-mr-reviewer/references/comment_templates.md- Conventional comment examples -
skills/gitlab-mr-reviewer/references/labels.md- Review workflow labels
Integration:
- Updated
index.jsto register the new skill in the skills catalog
Key Features
1. Conventional Comments Format
Implements the Conventional Comments standard:
-
suggestion:- Propose improvements -
issue:- Identify problems requiring fixes -
question:- Request clarification -
thought:- Share considerations -
nitpick:- Minor style preferences -
praise:- Highlight good work
2. Multi-Role Support
- Author Self-Review: Pre-submission checklist and validation
- Reviewer: Focus on correctness, quality, and testing
- Maintainer: Architecture fit and cross-domain consistency
3. Comprehensive Review Areas
- Code Quality: Logic, DRY principles, naming, abstractions
- Security: Input validation, authorization, sensitive data handling
- Performance: Query efficiency, caching, N+1 prevention
- Database: Migration reversibility, query performance, indexing
- Tests: Coverage, edge cases, isolation
- Documentation: Comments, API docs, changelog entries
4. Pipeline Analysis
- Check pipeline status
- Analyze failed jobs
- Review job logs
- Identify failure patterns
5. Suggested Code Changes
Format code suggestions that can be applied directly in GitLab UI:
suggestion: Extract repeated logic
```suggestion
def validate_user_permissions(user, resource)
return false unless user.active?
user.can_access?(resource)
end
```
6. Workflow Label Management
Suggest appropriate workflow labels based on review outcome:
-
workflow::in review- Under active review -
workflow::blocked- Changes required - Ready for maintainer approval
Why This Change
This skill completes the GitLab workflow automation trilogy:
-
✅ gitlab-issue-epic-creator - Create issues and epics -
✅ gitlab-mr-creator - Create merge requests -
✅ gitlab-mr-reviewer - Review merge requests (NEW)
Together, these skills provide end-to-end AI assistance for the entire GitLab development workflow, from issue creation through code review and merge.
Related Issues
Partially implements #2 (closed)
Testing
The skill has been structured following the same pattern as existing skills:
- Clear workflow documentation
- Comprehensive reference materials
- Practical examples and templates
- GitLab API tool integration
Documentation
All documentation is included in the skill's reference files:
- API usage examples
- Review guidelines and checklists
- Comment formatting templates
- Label conventions
Checklist
-
New skill files added under
skills/gitlab-mr-reviewer/ -
Skill registered in
index.js - Comprehensive reference documentation included
- Follows existing skill structure and patterns
- Related issue linked
- Commit message follows guidelines