CRITICAL: Source code provided contains only Android/frontend build configs and package.json files. NO tool implementations are visible. The 10 tools (cogcor, spotify, discord, lovense, telegram, biometrics, video, nanobanana, notion, catalogue) are referenced in src/index.ts as import statements and registration calls, but their actual definitions, schemas, descriptions, parameters, are not present in the provided source. Per hard scoring rules: 'If you cannot see the actual tool definition in the source (only inferred): cap that tool's overall at 50.' Since ALL 10 tools are inferred-only, each caps at 50. However, the inferred tools have ZERO visible descriptions ('details not available in provided source') and ZERO visible schemas.

• SUBMIT COMPLETE SOURCE. Include all src/tools/*.ts files so tool definitions, schemas, descriptions, and parameter types can be evaluated.
• RENAME TOOLS with verb_noun convention. E.g., 'get_spotify_track', 'search_spotify_playlist', 'send_discord_message', 'get_telegram_chat_history' instead of generic '_tool_1' suffixes.
• ADD DESCRIPTIONS to every tool (10 - 1024 chars). State WHAT it does, WHEN to use it, and any prerequisites. Example: 'Send a direct message to a Discord user. Requires user ID or username. Returns message ID and timestamp.'
• DEFINE INPUT SCHEMAS with Zod or JSON Schema. Every parameter must have a type (string, number, enum, etc.), description, and constraints (min/max, pattern, enum values).
• DOCUMENT OUTPUT SCHEMAS. Specify what fields are returned, their types, and which are needed for downstream tool chaining (e.g., if a search returns user_id, ensure update tools accept user_id).

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

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