clean-architecture
Clean Architecture principles for maintainable, testable applications. Use when designing application structure or refactoring.
When & Why to Use This Skill
This Claude skill provides a comprehensive framework for implementing Clean Architecture principles, enabling the creation of maintainable, testable, and scalable software. It guides developers in decoupling core business logic from external frameworks, databases, and UI components through a strict dependency rule and layered structure, ensuring long-term project health and technical flexibility.
Use Cases
- Initial System Design: Establishing a robust folder structure and defining clear layer boundaries for new enterprise-grade applications.
- Legacy Code Refactoring: Restructuring 'Big Ball of Mud' or monolithic codebases into decoupled modules to reduce technical debt and improve maintainability.
- Technology Migration: Designing applications where core business rules are independent of specific web frameworks or databases, allowing for easier infrastructure swaps.
- Enhanced Testing Strategy: Implementing a structure that facilitates high-coverage unit testing of domain logic without the need for complex mocks or I/O operations.
| name | clean-architecture |
|---|---|
| description | Clean Architecture principles for maintainable, testable applications. Use when designing application structure or refactoring. |
Clean Architecture
Layer Structure
┌─────────────────────────────────────────────┐
│ FRAMEWORKS & DRIVERS (Web, DB, External) │
│ ┌─────────────────────────────────────────┐ │
│ │ INTERFACE ADAPTERS (Controllers, Repos) │ │
│ │ ┌─────────────────────────────────────┐ │ │
│ │ │ APPLICATION (Use Cases) │ │ │
│ │ │ ┌─────────────────────────────────┐ │ │ │
│ │ │ │ DOMAIN (Entities) │ │ │ │
│ │ │ └─────────────────────────────────┘ │ │ │
│ │ └─────────────────────────────────────┘ │ │
│ └─────────────────────────────────────────┘ │
└─────────────────────────────────────────────┘
Dependency Rule
Dependencies point inward only.
- Domain knows nothing about outer layers
- Use Cases know Domain, nothing else
- Adapters know Use Cases + Domain
- Frameworks know everything
Layers
Domain (Entities)
- Business objects with behavior
- Validation rules
- No framework dependencies
- No I/O operations
Application (Use Cases)
- Orchestrates domain objects
- One use case = one action
- Defines repository interfaces
- Input/Output DTOs
Interface Adapters
- Controllers: HTTP → Use Case
- Repositories: Use Case → Database
- Presenters: Use Case → Response format
Frameworks & Drivers
- Web framework (Express, FastAPI, Fiber)
- Database clients
- External APIs
- Configuration
Folder Structure
src/
├── domain/
│ └── entities/
├── application/
│ ├── use-cases/
│ └── interfaces/ # Repository interfaces
├── infrastructure/
│ ├── repositories/ # Interface implementations
│ └── external/ # API clients
└── interfaces/
├── http/ # Controllers, routes
└── cli/ # CLI handlers
Key Patterns
Dependency Injection
UseCase depends on RepositoryInterface (not implementation)
Controller injects concrete Repository into UseCase
Use Case Structure
1. Validate input
2. Load entities via repository
3. Execute domain logic
4. Persist via repository
5. Return output DTO
Repository Interface
Defined in: application/interfaces/
Implemented in: infrastructure/repositories/
When to Apply
Use Clean Architecture when:
- Application will grow/evolve
- Multiple entry points (API, CLI, queue)
- Team needs clear boundaries
- Testing is priority
Skip for:
- Simple CRUD apps
- Prototypes/MVPs
- Scripts/tools
Common Mistakes
- Entities depending on ORM/framework
- Use cases calling HTTP directly
- Business logic in controllers
- Skipping the interface layer
- Over-engineering simple apps
Testing Strategy
| Layer | Test Type | Mocks |
|---|---|---|
| Domain | Unit | None |
| Use Cases | Unit | Repositories |
| Adapters | Integration | External services |
| E2E | System | None |