Three tools with clear action verbs and reasonable descriptions (avg 156 chars), but significant gaps in schema completeness and parameter documentation. search_tools and call_tool have documented input schemas with type info, but output schemas are undocumented. load_mcp_server lacks parameter descriptions entirely. No error handling guidance, no enum constraints, no parameter validation rules stated. Tool composition is sound (search → call → load pattern), but parameter descriptions are sparse or missing.

• Document output schemas for all three tools. For search_tools, specify: returns array of {tool: {qualified_name: string, description: string, input_schema: object}, score: number}. For call_tool and load_mcp_server, document success and error response structures.
• Add parameter descriptions to load_mcp_server: 'server_name' should state format (alphanumeric, unique, used as prefix in tool names); 'url' should specify protocol (https required), timeout, and validation rules.
• Add error handling guidance to call_tool description: 'If the tool is not found, use search_tools to rediscover. If auth fails, ask user to update tokens file. If arguments are invalid, check the tool schema from search results.'
• Constrain search_tools top_k: add min=1, max=50 to the parameter schema and state in description: 'Limit results to 1 - 50. Higher values return more options but increase token cost.'
• Add validation rules to call_tool arguments description: 'Must match the tool's input_schema from search_tools results. Missing required fields or type mismatches will cause the call to fail.'