Single tool 'recent_conversations' has a well-structured schema with proper types and descriptions for both parameters (limit, offset). Tool description is present and informative (Chinese + English security notice), explaining the read-only nature and untrusted data warning. However, the description lacks explicit guidance on WHEN to use this tool vs alternatives, and output schema is not documented in the visible code. Parameter descriptions are adequate but could be more explicit about constraints (e.g., 'limit must be 10-50, default 10'). No error handling guidance visible. Security notice is good but should be in parameter descriptions, not just tool description.

• Document the output schema explicitly: 'Returns an array of conversation objects, each with: id (string), title (string), timestamp (ISO 8601 string). Example: [{"id": "conv-123", "title": "Project Discussion", "timestamp": "2024-01-15T10:30:00Z"}]'
• Enhance parameter descriptions with explicit constraints: 'limit: Number of conversations to return. Must be 1-50; default 10. Values outside this range are clamped.' and 'offset: Pagination offset (0-1000000). Returns empty array if offset >= total conversations.'
• Add error handling guidance to tool description: 'If RikkaHub is unavailable, returns error with retry guidance. If authentication fails, check RIKKAHUB_WEB_PASSWORD. Network timeouts (>15s) indicate connectivity issues.'
• Rename tool to 'list_recent_conversations' to follow verb-first naming convention and improve LLM discoverability.
• Add a note about pagination: 'Use limit and offset to paginate through large conversation histories. Combine with timestamps to detect new conversations since last call.'

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

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