Single tool 'get-projects' has a basic description but critically lacks parameter descriptions and output schema documentation. The input schema shows a 'name' parameter with type 'string' and optional flag, but no description of what filtering behavior it enables or expected format. No output schema is visible in the source code. The tool definition appears incomplete and would not meet production standards for LLM-driven agent selection.

• Add a detailed output schema showing the structure of returned projects (e.g., {id: string, name: string, description: string, created_at: ISO8601, owner_id: string}). Include all fields needed for downstream tool chaining.
• Expand tool description to 100-150 chars: 'Retrieve projects filtered by optional name. Returns up to 50 projects with IDs, names, and metadata. Use this to discover available projects before creating tasks or accessing project details.'
• Add description to 'name' parameter: 'Optional substring filter for project name (case-insensitive). If omitted, returns all projects accessible to the caller.'
• Add pagination support: include 'limit' (1-100, default 20) and 'offset' (default 0) parameters. Return total_count in response so LLM knows if more results exist.
• Document error cases: 'Returns empty list if no projects match. Returns 403 if caller lacks list:projects permission. Returns 500 with retry guidance if database is unavailable.'

Inferred effective spec: <=2025-11-25.

• Relies on Stateful initialize / Mcp-Session-Id (removed; protocol is stateless) - make each request self-contained