Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
CartoGopher parses a codebase into a compact graph and exposes 30+ MCP tools for symbol lookup, API tracing, architecture mapping, and AST-aware search. Setup involves getting a 7-day trial API key and configuring the MCP server for your AI tool.
CartoGopher uses a pre-parsed code graph instead of raw file reads, cutting token usage by 94% compared to reading files directly.
When to use it
- Tracing an API endpoint from frontend to database
- Finding where a function or type is defined and used across a large repo
- Understanding project architecture before making structural changes
- Searching code semantically without grep-style text matching
- Analyzing React or Vue component props, hooks, and store usage
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: set up CartoGopher for my project
- 1Ask the user for their email address
- 2POST to cartogopher.com/api/agent-trial with the email to get a trial API key
- 3Download and extract the platform-appropriate MCP server bundle to ~/.cartogopher/mcp-server
- 4Add CARTOGOPHER_API_KEY to the user's shell profile and run npm install in the server directory
- 5Write the MCP server config for the user's AI tool (Claude Code, Cursor, or OpenClaw) with the API key
CartoGopher MCP server is configured and verified working via the shake tool on the user's project
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.