Server has 5 tools with clear verb-noun naming (add_todo, list_todos, get_todo, complete_todo, delete_todo). All tools have descriptions (10-50 chars) and visible input schemas with type definitions. However, descriptions are minimal (under 50 chars), lacking context on WHEN to use each tool or what happens on error. Parameter descriptions are present but terse. Output schemas are not documented, LLMs cannot predict response structure. Error handling returns dict with 'error' or 'message' keys but provides no recovery guidance. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels (WRITE, DESTRUCTIVE). Missing pagination guidance for list_todos despite potential for growth.

• Add output schema documentation to each tool. Example for add_todo: 'Returns {"message": string, "todo": {"id": int, "title": string, "description": string, "completed": bool, "created_at": ISO8601 string}}'
• Expand tool descriptions to 80-150 chars with WHEN/WHY context. Example: 'Add a new todo item. Use this to create tasks. Returns the new todo with auto-generated ID. If title is empty, returns error, validate before calling.'
• Add tool annotations via FastMCP. Mark delete_todo with destructiveHint=true, complete_todo with idempotentHint=true, list_todos with readOnlyHint=true.
• Enhance error messages with recovery paths. Instead of 'No todo found with id 5', return 'No todo found with id 5. Call list_todos() to see available IDs (1-3).'
• Add pagination to list_todos: accept optional 'limit' (default 20, max 100) and 'offset' (default 0) parameters. Return {"count": int, "total": int, "todos": [...], "has_more": bool}.