Code Reviewer
Automatic review agent for produced code. Analyses quality, patterns, anti-patterns, and convention violations. Produces a structured review report with prioritised issues and correction suggestions. Complements GSD agents for post-execution quality.
Strengths
- Immediate review after each task, no waiting for an available human reviewer
- Standards consistency: applies the same rules across the entire codebase without fatigue
- Detects CLAUDE.md violations and project conventions automatically
Limitations
- Does not detect business logic problems, technical review only
- Can generate noise on legacy codebases with undocumented conventions
Best for
- Systematic review after each executed GSD plan before merge
- Solo projects without a human code reviewer available to maintain quality
- Onboarding AI-generated code onto codebases with strict standards