| name | architecture-review |
| description | Reviews the system architecture for structural issues, dependency cycles, layer violations, scalability concerns, and single points of failure. Use for periodic health checks or before major new features. |
| argument-hint | [component, module, or 'full' for entire project] |
| allowed-tools | Read, Grep, Glob, Bash |
Architecture Review
Project Structure
!find . -type f -not -path "*/node_modules/*" -not -path "*/.git/*" -not -path "*/vendor/*" -not -path "*/__pycache__/*" -not -path "*/target/*" -not -path "*/bin/*" -not -path "*/obj/*" -not -path "*/.next/*" -not -path "*/dist/*" -not -path "*/build/*" 2>/dev/null | head -80
Directory Layout
!tree -L 3 -I "node_modules|.git|vendor|__pycache__|target|bin|obj|.next|dist|build" 2>/dev/null || find . -type d -not -path "*/node_modules/*" -not -path "*/.git/*" -not -path "*/vendor/*" -maxdepth 3 2>/dev/null
Package/Module Config
!cat package.json go.mod Cargo.toml pyproject.toml requirements.txt *.csproj pom.xml 2>/dev/null | head -60
Review Areas
1. Project Structure & Organization
- Is the directory layout clear and consistent?
- Does it follow established conventions for the stack? (e.g., src/, app/, lib/, pkg/)
- Are related files co-located or scattered?
- Is there a clear entry point?
2. Dependency Architecture
- Map the dependency graph between major modules/packages
- Identify circular dependencies — modules that import each other
- Check for layer violations — e.g., presentation layer importing directly from data layer
- Look for god modules — modules that everything depends on
- Assess coupling: can modules be changed independently?
3. Separation of Concerns
- Is business logic separated from framework/transport code?
- Are side effects (I/O, network, database) isolated at boundaries?
- Is configuration separated from code?
- Are there clear interfaces between layers?
4. Scalability Considerations
- Stateless vs stateful components — which components hold state?
- Are there bottlenecks that would fail under load?
- Database query patterns — are they efficient at scale?
- Caching strategy — what is cached, where, and when invalidated?
- Background job handling — how are long-running tasks managed?
5. Reliability & Resilience
- Single points of failure: What happens if any component goes down?
- Error boundaries: Where do errors propagate, and are they caught?
- Retry logic: For external service calls, is there retry with backoff?
- Graceful degradation: Does the system handle partial failures?
- Health checks: Are there endpoints or mechanisms to monitor health?
6. Security Architecture
- Authentication and authorization boundaries
- Where is input validation enforced?
- How are secrets managed?
- API rate limiting and abuse prevention
Output Format
Architecture Diagram (ASCII)
┌─────────┐ ┌─────────┐ ┌──────────┐
│ Frontend │────▶│ API │────▶│ Database │
└─────────┘ └─────────┘ └──────────┘
│
▼
┌───────────┐
│ External │
│ Services │
└───────────┘
Findings Table
| # | Area | Finding | Severity | Recommendation |
|---|
Top Recommendations
Prioritized list of the most impactful architectural improvements.
Overall Health
Rate: Healthy / Needs Attention / At Risk — with reasoning.