ManSit MCP server has basic tool definitions with schemas and descriptions present, but falls short of production quality. All 6 tools have descriptions (10-60 chars, below the 194-char baseline) and input schemas with type definitions. However, descriptions are generic and lack actionable context for LLM selection. No output schemas documented. Parameter descriptions are minimal. No error recovery guidance. No tool annotations (readOnlyHint, etc.). Tools are read-only but this is not explicitly declared in descriptions. Missing pagination support despite list_conversations and list_vault_files returning unbounded results. No parameter constraints (enums, ranges, patterns) despite accepting free-form strings.

• Expand tool descriptions to 100-200 chars. Example: 'search_memories: Search conversation history by keyword. Returns matching messages with timestamps. Use this to find past context; use get_conversation if you have a specific session ID.'
• Document output schemas for all tools. Specify that search_memories returns {query, count, results: [{key, data}]}, list_conversations returns {count, conversations: [{sessionId, timestamp, messageCount}]}, etc.
• Add pagination to list_conversations and list_vault_files: add 'offset' (default 0) and 'limit' (default 20, max 100) parameters; return 'total_count' and 'has_more' in response.
• Add tool annotations: set readOnlyHint=true on all 6 tools to signal safety to LLMs and enable optimizations like caching.
• Add parameter constraints: specify that 'query' accepts free-form text (no regex), 'limit' ranges 1-100, 'directory' defaults to root (''), and 'sessionId' is a UUID or opaque string.