Project Codemap (command)
Contributed by Rohit Ghumare (awesome-claude-code-toolkit)
Source: original file · added 2026-10-29
Update the project codemap to reflect the current state of the codebase.
Steps
1. Scan Project Structure
- Run
find . -type f -not -path '*/node_modules/*' -not -path '*/.git/*' -not -path '*/dist/*' -not -path '*/__pycache__/*'to get the file tree. - Identify the top-level directories and their purposes.
- Note the primary language(s) and framework(s) from config files (package.json, pyproject.toml, go.mod, Cargo.toml).
2. Identify Key Modules
For each major directory:
- Purpose: What this module is responsible for.
- Entry point: The main file that exports or bootstraps the module.
- Key files: The 3-5 most important files and what they contain.
- Dependencies: Which other modules this one depends on.
3. Map Data Flow
- Trace the request lifecycle: entry point -> routing -> business logic -> data access -> response.
- Identify shared state: databases, caches, message queues, shared config.
- Note async boundaries: background jobs, event handlers, webhooks.
4. Document Key Patterns
- Architecture pattern (MVC, layered, hexagonal, microservices).
- State management approach.
- Error handling strategy.
- Testing strategy and test location conventions.
5. Output Format
# Project Codemap
## Structure
src/
api/ - REST API handlers and middleware
services/ - Business logic layer
models/ - Database models and schemas
utils/ - Shared utilities
## Key Files
- src/api/server.ts - Express app setup, middleware chain
- src/services/auth.ts - Authentication and JWT handling
- src/models/user.ts - User model with validation
## Data Flow
Request -> Middleware (auth, validation) -> Handler -> Service -> Repository -> Database
## Patterns
- Dependency injection via constructor parameters
- Repository pattern for data access
- Result type for error handling (no thrown exceptions in services)Rules
- Keep the codemap concise. It should fit in one screen for small projects, two for large ones.
- Focus on orientation, not exhaustive listing. A new developer should understand the project structure in 2 minutes.
- Update the codemap in CLAUDE.md or a dedicated CODEMAP.md at the project root.
- Include the build and test commands.
Original source file (unmodified)
Update the project codemap to reflect the current state of the codebase. ## Steps ### 1. Scan Project Structure - Run `find . -type f -not -path '*/node_modules/*' -not -path '*/.git/*' -not -path '*/dist/*' -not -path '*/__pycache__/*'` to get the file tree. - Identify the top-level directories and their purposes. - Note the primary language(s) and framework(s) from config files (package.json, pyproject.toml, go.mod, Cargo.toml). ### 2. Identify Key Modules For each major directory: - **Purpose**: What this module is responsible for. - **Entry point**: The main file that exports or bootstraps the module. - **Key files**: The 3-5 most important files and what they contain. - **Dependencies**: Which other modules this one depends on. ### 3. Map Data Flow - Trace the request lifecycle: entry point -> routing -> business logic -> data access -> response. - Identify shared state: databases, caches, message queues, shared config. - Note async boundaries: background jobs, event handlers, webhooks. ### 4. Document Key Patterns - Architecture pattern (MVC, layered, hexagonal, microservices). - State management approach. - Error handling strategy. - Testing strategy and test location conventions. ### 5. Output Format ```markdown # Project Codemap ## Structure src/ api/ - REST API handlers and middleware services/ - Business logic layer models/ - Database models and schemas utils/ - Shared utilities ## Key Files - src/api/server.ts - Express app setup, middleware chain - src/services/auth.ts - Authentication and JWT handling - src/models/user.ts - User model with validation ## Data Flow Request -> Middleware (auth, validation) -> Handler -> Service -> Repository -> Database ## Patterns - Dependency injection via constructor parameters - Repository pattern for data access - Result type for error handling (no thrown exceptions in services) ``` ## Rules - Keep the codemap concise. It should fit in one screen for small projects, two for large ones. - Focus on orientation, not exhaustive listing. A new developer should understand the project structure in 2 minutes. - Update the codemap in CLAUDE.md or a dedicated CODEMAP.md at the project root. - Include the build and test commands.