Four tools with complete schemas and descriptions. Naming follows verb_noun pattern (search_, get_, list_). Descriptions are substantive (150-250 chars) and explain WHEN to use each tool. However, output schemas are undocumented, LLMs cannot predict response structure. Parameter descriptions lack format constraints (e.g., 'episode' accepts 'Brian Chesky.txt' but no validation rules stated). No error handling guidance. Missing tool annotations (readOnlyHint, idempotentHint). Composition is sound: tools chain logically (search → get_chapter → get_full_transcript).

• Document output schemas for all tools. Example for search_lenny: 'Returns array of objects with fields: type (string), episode_file (string), relevance_score (float 0-1), guest (string), expertise_tags (array), topic_id (string), text (string), lesson (string), tags (array).'
• Add format constraints to 'episode' parameter: 'Episode filename as returned by list_episodes (e.g., "Brian Chesky.txt"). Must match an available episode or tool will return error.'
• Add error recovery hints to get_chapter and get_full_transcript descriptions: 'If episode not found, call list_episodes() to discover available episodes. If topic_id invalid, call get_full_transcript() to see all topic IDs for that episode.'
• Declare tool annotations in inputSchema. Add 'readOnlyHint': true to all four tools (they do not modify state). Example: Tool(..., inputSchema={..., 'readOnlyHint': true}).
• Specify minimum for top_k: 'Number of results to return (default: 5, min: 1, max: 20). Must be a positive integer.'