Six tools with basic verb-noun naming (get_, list_, search_, compare_) and present descriptions. However, descriptions are minimal (10-60 chars), parameter descriptions lack detail, and no output schemas are documented. Input schemas exist but lack depth, most parameters have type and description, but no enums, ranges, or format constraints. No error handling guidance visible. Tools are read-only (low risk) but lack the structure needed for reliable LLM invocation. Composition is reasonable (single responsibility per tool), but parameter design does not match chat data model (e.g., 'pokemonId' accepts both ID and name but lacks guidance on format).

• Document output schemas for all six tools. Specify which fields are returned, their types, and which are required. E.g., get_pokemon should return {id, name, types: string[], stats: {hp, attack, ...}, imageUrl}.
• Expand tool descriptions to 100-200 chars. Answer: What does it do? When should the LLM call it? What does it return? E.g., 'Get detailed Pokémon info (types, stats, abilities, evolution chain) by ID or name. Use this to answer questions about a specific species.'
• Add format/constraint guidance to parameters. E.g., 'pokemonId: Pokémon identifier (numeric ID 1-1025 or English name, case-insensitive)' and 'query: Search term (1-40 chars, supports partial names and type names).'
• Add pagination to list_my_collection. Accept limit (default 20, max 100) and offset/cursor. Return total count and next_cursor so LLM can fetch more if needed.
• Add error recovery hints. E.g., 'If Pokémon not found, try search_pokemon() with a partial name.' Document what empty results mean (no matches vs. collection empty).