Critical failures across all 14 tools. Descriptions are present but generic (10-50 chars), below the 50-200 char baseline for LLM-optimized docs. No parameter descriptions exist because no parameters are defined. Tools lack context on what IDs they accept, what data they return, pagination support, or error recovery paths. The server reads from WHOOP API but exposes no schema documentation of responses. This is a skeleton implementation with no production-grade tool definitions.

• Add input schemas to all 14 tools. For get_* tools, define required parameters (e.g., cycle_id: string, sleep_id: string). For list_* tools, add limit (1-100, default 20), offset (default 0), and date_range parameters.
• Expand descriptions to 80-150 characters. Example: 'Get the user's WHOOP profile including name, email, and account creation date. Call this first to verify authentication and retrieve the user_id for other queries.'
• Document output schemas for each tool. Specify which fields are returned (e.g., cycle_id, start_time, end_time, strain_score) so LLMs can chain tools and extract IDs for follow-up calls.
• Add parameter descriptions. For cycle_id, write 'The unique WHOOP cycle identifier (UUID format). Retrieve from list_cycles() or get_today_snapshot().' This guides LLMs on how to obtain required values.
• Implement error handling with recovery guidance. E.g., 'Cycle not found. Call list_cycles() to see available cycles for the user.' This prevents dead-end failures and guides agent planning.

Inferred effective spec: 2026-07-28+.

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