Code Quality
Code quality analysis examines your codebase through four specialized reviewers that run in parallel: security, quality, performance, and architecture.
Four-reviewer architecture
The CodeReviewWorkflow coordinates four subagents, each focusing on a specific domain:
- Security reviewer — Identifies vulnerabilities, unsafe patterns, and potential attack vectors
- Quality reviewer — Catches style violations, logical errors, and maintainability issues
- Performance reviewer — Spots inefficient algorithms, memory leaks, and scalability bottlenecks
- Architecture reviewer — Evaluates design patterns, coupling, and structural health
Each reviewer analyzes the same code independently, then their findings are synthesized into a unified report with an overall health score (0-100).
Unified reporting format
Instead of reading four separate outputs, you get a single structured report:
| Section | Content | Source |
|---|---|---|
| Summary | Overall health score and executive summary | Synthesized from all reviewers |
| Security | Vulnerabilities and safety issues | Security reviewer |
| Quality | Style, bugs, and maintainability | Quality reviewer |
| Performance | Efficiency and scalability concerns | Performance reviewer |
| Architecture | Design and structural analysis | Architecture reviewer |
| Suggestions | Actionable next steps by priority | Cross-reviewer synthesis |
When specialized review matters
This four-reviewer approach catches issues that single-purpose tools miss. A function might pass your linter (style), have no obvious bugs (quality), but create a timing attack vulnerability (security) or scale poorly under load (performance). Running all four perspectives simultaneously gives you comprehensive coverage without the overhead of managing separate tools.