Wanikani-MCP has 6 tools with basic verb-noun naming (get_*, find_*) and reasonable descriptions (50-150 chars). However, critical gaps undermine quality: most parameters lack descriptions entirely, input schemas are minimal or missing type constraints, and output schemas are undocumented. Tool descriptions are present but generic, they state WHAT but not WHEN or WHY to use each tool. No error handling guidance, no pagination support despite tools returning lists, and no indication of which fields downstream tools need. The server follows a single domain (Japanese language learning) with coherent naming, but falls short of production-grade definition quality.

• Document output schemas for all tools. Specify return type (object/array), required fields, and field types. Example: 'Returns array of {subject_id: integer, character: string, level: integer, onyomi: string[], kunyomi: string[]}'.
• Add parameter descriptions for all inputs. For 'characters' array: 'Array of 1-50 kanji characters (e.g., ["漢", "字"]). Invalid characters are silently skipped.' For 'phrase': 'Vocabulary phrase or slug (e.g., "漢字" or "kanji"). Case-insensitive.'
• Implement pagination for list-returning tools. Add 'limit' (default 20, max 100) and 'offset' (default 0) parameters. Return total_count and has_more in response.
• Add error handling guidance to descriptions. E.g., 'Returns 404 if subject_id not found. Try get_level_info() to see available subject ranges for your level.'
• Document tool composition and chaining. Add a note: 'Use get_kanji_by_character() to find kanji in a sentence, then pass returned subject_ids to get_subject_by_id() for detailed readings and mnemonics.'

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

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