Use when adding new error messages to React, or seeing "unknown error code" warnings.
npx skills add DonggangChen/antigravity-agentic-skills --skill "peer_review"
Install specific skill from multi-skill repository
# Description
Academic/technical document review, methodology evaluation. ⚠️ Use for document/research. For code review → code-review.
# SKILL.md
name: peer_review
router_kit: FullStackKit
description: Academic/technical document review, methodology evaluation. ⚠️ Use for document/research. For code review → code-review.
metadata:
skillport:
category: research
tags: [architecture, automation, best practices, clean code, coding, collaboration, compliance, debugging, design patterns, development, documentation, efficiency, git, optimization, peer review, productivity, programming, project management, quality assurance, refactoring, software engineering, standards, testing, utilities, version control, workflow] - quality
📝 Peer Review
Academic and technical peer review methodology guide.
📋 Review Framework
Evaluation Areas
| Area | Questions |
|---|---|
| Clarity | Is it open and understandable? |
| Methodology | Is the method appropriate? |
| Validity | Are results valid? |
| Originality | Is there original contribution? |
| Completeness | Is anything missing? |
🔍 Code Review Checklist
## Functionality
- [ ] Does code work as expected?
- [ ] Are edge cases handled?
- [ ] Is error handling sufficient?
## Code Quality
- [ ] Is DRY principle applied?
- [ ] Is naming convention consistent?
- [ ] Are comments sufficient?
## Security
- [ ] Is there input validation?
- [ ] Is there SQL injection risk?
- [ ] Is sensitive data protected?
## Performance
- [ ] Is there unnecessary operation?
- [ ] Is there memory leak risk?
📄 Document Review Template
## Review Summary
**Document:** [Document Name]
**Reviewer:** [Name]
**Date:** [Date]
## Overall Assessment
[General assessment - 1-2 paragraphs]
## Strengths
1. ...
2. ...
## Areas for Improvement
1. ...
2. ...
## Specific Comments
| Section | Comment | Severity |
| ------- | ------- | ----------- |
| ... | ... | Major/Minor |
## Recommendation
[ ] Accept
[ ] Minor Revisions
[ ] Major Revisions
[ ] Reject
💬 Constructive Feedback
Good Feedback
✅ "This function might error in X condition.
Can you consider adding Try-catch?"
✅ "Nice implementation! A suggestion:
It would be more readable if this method is extracted."
To Avoid
❌ "This is wrong"
❌ "Why did you do this?"
❌ "I wouldn't do it like this"
Peer Review v1.1 - Enhanced
🔄 Workflow
Source: Conventional Comments & Google Engineering Practices
Phase 1: Preparation (Reviewee)
- [ ] Self-Review: Read yourself before pushing code, clean unnecessary logs.
- [ ] Context: Answer "What?", "Why?" and "How to Test?" questions in PR description.
- [ ] Scope: Keep change in manageable size (<400 lines preferably).
Phase 2: Review Process (Reviewer)
- [ ] Clarity: Does code explain what it does? Are variable names descriptive?
- [ ] Security: Is user input sanitized? Is there Auth check?
- [ ] Performance: Are there unnecessary loops or N+1 queries?
- [ ] Tone: Use tags like
nit:,suggestion:,blocking:in comments (Conventional Comments). Use "Code" language instead of "You" language.
Phase 3: Approval & Merge
- [ ] CI Checks: Are all tests and lint checks green?
- [ ] Resolution: Are all critical comments (blocking) resolved?
- [ ] Squash & Merge: Combine commits for history cleanliness.
Checkpoints
| Phase | Verification |
|---|---|
| 1 | Does PR description include screenshot or video (if Frontend)? |
| 2 | Did Reviewer pull code to local and run it (Complex changes)? |
| 3 | Is Feedback constructive or just critical? |
# Supported AI Coding Agents
This skill is compatible with the SKILL.md standard and works with all major AI coding agents:
Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.