Four tools with basic descriptions but significant schema and parameter documentation gaps. Tool names are action-oriented (search, skills, compositions, leaderboards) but descriptions lack depth on when/why to use each tool. Input schemas visible but parameters lack detailed descriptions explaining constraints, formats, or dependencies. No output schemas documented. Error handling guidance absent. Compositions tool explicitly returns not-found for all queries (MVP stub), reducing practical utility. No tool annotations (readOnlyHint, etc.) despite all being read-only operations.

• Document output schemas for all four tools. Example for search: 'Returns {results: [{skillId, name, confidence, description}], total, nextCursor}' so LLMs know what fields to extract.
• Expand parameter descriptions with constraints. E.g., 'query (string, 1-256 chars): Search term for skill name or description. Supports partial matches.' and 'limit (number, 1-100, default 20): Max results per page.'
• Remove or replace the compositions tool. If MVP, document it as 'Not yet implemented, returns 404 for all queries. Check back in v2.0.' If functional, implement it fully.
• Add readOnlyHint: true to all four tool definitions to signal safety to agents and enable optimizations (e.g., parallel execution, no transaction wrapping).
• Add error recovery guidance to each tool description. E.g., 'If skillId not found, try search() with a partial name. Available skill categories: [list]' to guide agent retry logic.

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

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry