Review Mode Skill | Agent Skills

Review Mode

Goal: Evaluate code quality, identify issues, and suggest improvements.

Process

Understand the code's purpose and language
Review against language-specific standards
Identify issues by severity
Suggest specific improvements
Highlight what's done well

Output Format

## REVIEW: [Component/Feature name]

**Scope:** [What was reviewed]
**Language:** [JS/Python/Java/Go/PHP/Ruby]
**Overall:** [Good / Needs Work / Critical Issues]

---

### Summary
| Category | Status |
|----------|--------|
| Functionality | OK / Warning / Error |
| Code Quality | OK / Warning / Error |
| Security | OK / Warning / Error |
| Performance | OK / Warning / Error |
| Maintainability | OK / Warning / Error |

---

### Issues Found

#### Critical
| Issue | Location | Suggestion |
|-------|----------|------------|
| [Description] | `file:line` | [How to fix] |

#### Important
| Issue | Location | Suggestion |
|-------|----------|------------|
| [Description] | `file:line` | [How to fix] |

#### Minor / Suggestions
| Issue | Location | Suggestion |
|-------|----------|------------|
| [Description] | `file:line` | [How to fix] |

---

### What's Done Well
- [Positive point 1]
- [Positive point 2]

---

### Recommended Actions
1. [ ] [Action 1] - Priority: High
2. [ ] [Action 2] - Priority: Medium
3. [ ] [Action 3] - Priority: Low

Review Checklist

Code Quality (All Languages)

[ ] No loose typing (no any, raw Object, interface{} abuse)
[ ] Meaningful names
[ ] No duplicate code
[ ] Small functions (< 50 lines)
[ ] Proper error handling

Security (All Languages)

[ ] No hardcoded secrets/credentials
[ ] Input validation present
[ ] SQL injection prevention (parameterized queries)
[ ] XSS prevention (output escaping)
[ ] Proper authentication/authorization
[ ] Sensitive data not logged

Performance

[ ] No unnecessary loops/iterations
[ ] No memory leaks (proper cleanup)
[ ] Efficient algorithms (avoid O(n²) when O(n) possible)
[ ] No N+1 queries

Maintainability

[ ] Code is self-documenting
[ ] Complex logic has comments explaining WHY
[ ] Follows project/language conventions
[ ] Easy to test (dependencies injectable)

Severity Levels

| Level | Description | Action Required | |-------|-------------|-----------------| | Critical | Security vulnerability, data loss, crash | Must fix before merge | | Important | Bug, bad practice, tech debt | Should fix soon | | Minor | Style, optimization, nice-to-have | Optional improvement |

Principles

| DON'T | DO | |-------|-----| | Only criticize | Balance with positive feedback | | Be vague ("this is bad") | Be specific with location and fix | | Focus on style only | Prioritize functionality and security | | Rewrite everything | Suggest minimal effective changes | | Skip context | Understand purpose before reviewing |

Agent Skills: Review Mode

Install this agent skill to your local

Skill Files