This interface is designed so agents can operate on repositories remotely, without cloning source code to the machine running the agent. The server exposes MCP capabilities in three groups:
GroupDescription
ToolsActive operations: project management, file editing, search, git, and command execution
ResourcesPassive read-only data: project tree, file content, metadata, and git status
PromptsPre-built workflow instructions: implement, analyze, review, generate commit, fix tests

​
Response envelope

All tool responses share a consistent envelope format. Successful operation: 
{
  "ok": true,
  "operation": "read_file",
  "project_id": "abc123",
  "data": {},
  "warnings": [],
  "errors": [],
  "request_id": "uuid",
  "duration_ms": 12
}
Failed operation: 
{
  "ok": false,
  "operation": "write_file",
  "error_code": "SECURITY_ERROR",
  "message": "Operation blocked for protected path",
  "details": {}
}

​
Where handlers are implemented

  • src/mcp/tools/* — tool handlers
  • src/mcp/resources/* — resource handlers
  • src/mcp/prompts/* — prompt handlers
Use the reference pages in this section to review the full capability surface.