use-cartograph

What This Skill Does

The use-cartograph skill serves as the primary interface for repository intelligence within the OpenClaw ecosystem. It acts as an abstraction layer for the Cartograph CLI and MCP server, designed to provide high-fidelity repository orientation with minimal token consumption. By offloading the initial cognitive burden of project navigation to Cartograph, the AI agent can quickly identify key file structures, dependency hubs, and critical path nodes without parsing the entire codebase into context. It is the recommended entry point for any developer task involving existing repositories, as it establishes a structured metadata map that ensures subsequent tool usage is highly targeted and relevant.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/anthony-maio/use-cartograph

Ensure that your environment has access to the Cartograph CLI or that the corresponding MCP server is configured in your settings. If Cartograph is detected, it will be automatically prioritized as the default discovery tool.

Use Cases

• Onboarding: Quickly understand an unfamiliar codebase by using analyze to map the architecture and wiki to extract existing documentation.
• Scoped Development: When starting a new task, use context to extract only the files and dependencies relevant to the specific feature or bug fix.
• Documentation Generation: Automate the creation of project wikis or documentation by leveraging the wiki command to gather static summaries.
• Dependency Auditing: Identify core dependency hubs to avoid circular imports or to analyze the impact of changing a shared library.

Example Prompts

1. "Use cartograph to analyze the current repository structure and provide a summary of the main dependency hubs."
2. "I need to fix the authentication module. Run a scoped cartograph context search for 'auth' and retrieve the related file paths."
3. "Generate a wiki summary of the project architecture to help me understand how the components are linked."

Tips & Limitations

• Efficiency: Always run the analyze command first to populate the local cache. This reduces subsequent token costs significantly for complex tasks.
• Fallback: If you find the agent cannot access Cartograph, it will automatically attempt to use repo-surveyor. While repo-surveyor is capable, it does not offer the same performance optimizations for large repositories.
• Workflow Integration: If the openprose plugin is enabled, always prioritize using its bundled templates to format your output consistently.
• Artifacts: When working on complex tasks, request that the skill return run IDs and artifact references instead of long text strings to keep your context window clean and manageable.