Server has 5 tools with basic descriptions and some parameter documentation, but significant gaps in schema completeness, parameter descriptions, and error handling. Tool names are action-oriented (get_*, positive). However, input schemas are incomplete or missing for several tools, descriptions lack LLM-optimization guidance, and no output schemas are documented. Error handling is minimal. The codebase shows two different implementations (server.py variants) with inconsistent tool signatures, suggesting incomplete refactoring. Security is good (SSRF hardening via mycelium-security), but definition quality falls short of production standards.

• Document output schema for each tool. Example for get_today_summary: 'Returns {date: str, productivity_pulse: int (0-100), total_hours: float, very_productive_hours: float, productive_hours: float, neutral_hours: float, distracting_hours: float, very_distracting_hours: float, total_productive: float, total_distracting: float}'. This enables LLMs to extract fields and chain tools.
• Add minimum constraint to get_productivity_trend days parameter: 'Number of days to look back (min: 1, default: 7, max: 14)'. Prevent LLMs from passing 0 or negative values.
• Enhance date_str parameter descriptions with format guidance: 'Date to query as ISO 8601 string (YYYY-MM-DD), or special values: "today", "yesterday". Defaults to today.'
• Implement error classification and recovery guidance. Example: 'If RescueTime API returns 401, check RESCUETIME_API_KEY env var. If 429 (rate limit), retry after 60s. If no data for date, RescueTime may not have synced yet, try yesterday or a recent date.'
• Consolidate server implementations. Remove root server.py or src/rescuetime_mcp/server.py variant. Ensure single canonical tool set with consistent signatures and descriptions.