Server has 8 tools with mostly complete schemas and descriptions, but significant gaps reduce quality. All tools have descriptions (10-50 chars baseline met), and input schemas are visible with type definitions. However, parameter descriptions are sparse or missing for many fields (e.g., 'limit' in list_lifelogs has null description). Output schemas are not documented, LLMs cannot predict response structure. Error handling exists but lacks recovery guidance. Tool naming is clear and verb-prefixed, but composition issues exist: list_lifelogs, get_paged_lifelogs, and get_lifelog overlap significantly, forcing LLMs to reason about which to call. manage_cache is a utility tool that doesn't fit the domain. No tool annotations (readOnlyHint/destructiveHint) despite clear risk classifications in metadata.

• Document output schemas for all tools. Specify that responses return {content: [{type: 'text', text: string}]} and clarify what structured data (if any) is embedded in the text field.
• Add descriptions to all parameters. For 'fields' array, document valid field names (e.g., 'title, time, id, speakers, markdown'). For 'limit', specify range (e.g., '1-100, default 20').
• Consolidate list_lifelogs and get_paged_lifelogs into a single tool with optional cursor parameter. Reduce cognitive load on LLM tool selection.
• Add tool annotations to schema registration: readOnlyHint=true for all READ_ONLY tools, destructiveHint=true for manage_cache clear actions. Use server.tool() with annotations parameter.
• Enhance error messages with recovery hints. E.g., 'Lifelog not found. Call list_lifelogs() to see available IDs, or check the date range with start/end parameters.'