Server has 4 tools with basic schemas and descriptions, but significant quality gaps. All tools have descriptions (10-50 chars) and Zod-based input schemas, but descriptions are too brief to guide LLM selection effectively. The 'pending-tasks' tool has a nonsensical parameter ('status: Not used but required for MCP protocol') indicating schema confusion. No output schemas documented. Error handling returns text responses without actionable recovery guidance. No tool annotations (readOnlyHint, destructiveHint). Parameter descriptions lack format/constraint details. Overall structure is present but falls short of production quality.

• Expand all tool descriptions to 100-200 chars. Include WHAT (action), WHEN (use case), and state-modification clarity. Example: 'Create a new task with a title and optional description. Use this to add items to your task list. Returns the created task ID and metadata. Idempotent: calling twice with the same title creates two separate tasks.'
• Remove 'pending-tasks' tool entirely. It duplicates list-tasks(status='pending'). Consolidate into a single, well-named tool with clear parameter semantics.
• Document output schemas for all tools. Define return type as a structured object: {id: string, title: string, description: string, completed: boolean, createdAt: string}. Include this in tool descriptions or a separate schema field.
• Add tool annotations: mark 'create-task' and 'complete-task' with destructiveHint=true; mark 'list-tasks' and 'pending-tasks' with readOnlyHint=true. This guides LLM reasoning about retry safety.
• Enhance error messages with recovery guidance. Example: 'Task with ID abc123 not found. Call list-tasks to see available task IDs, or create a new task with create-task.'