Server has 6 tools with complete input schemas and basic descriptions. All tools follow verb_noun naming (get_*, list_*) and accept human-friendly parameters (Pokemon names, move names). However, descriptions are generic and lack context about when to use each tool or what distinguishes them. Parameter descriptions are minimal (e.g., 'Pokemon name' with no format guidance). Output schemas are not documented, LLMs cannot predict response structure. Error handling returns bare error strings without recovery guidance. No tool annotations (readOnlyHint, idempotentHint) despite all being read-only operations.

• Document output schemas for all tools. Example for get_pokemon: 'Returns {name: string, types: [string], stats: {hp: int, attack: int, ...}, height: int, weight: int}'.
• Expand tool descriptions to 100-150 chars explaining WHEN to use each tool. E.g., 'get_pokemon: Use when you need stats, types, height, weight. For moves only, use get_pokemon_moves instead.'
• Add format guidance to parameter descriptions. E.g., 'Pokemon name (lowercase, e.g. pikachu, charizard, mr-mime)'.
• Enhance error messages with recovery hints. E.g., 'Pokemon "pikachu" not found. Try list_pokemon() to see available names, or check spelling.'
• Add readOnlyHint: true annotation to all 6 tools to signal they are safe to retry and have no side effects.

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

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