This server has severe definition quality issues. Most tools lack input schemas entirely (web_search, read_file, write_file, list_directory have no visible schemas). Descriptions are minimal or generic ('Read file operations', 'Write file operations'). Only 3 of 8 tools have documented input schemas. No output schemas are documented. No error handling guidance. No parameter descriptions for most tools. The codebase shows a custom agent framework, not an MCP server, tool definitions are inferred from decorator usage rather than explicit MCP registration.

• Add complete input schemas to all 8 tools. For web_search: {query: string, limit: integer (1-50), offset: integer}. For read_file: {path: string, encoding: string (utf-8|ascii)}. For write_file: {path: string, content: string, overwrite: boolean}.
• Expand descriptions to 100-200 chars. E.g., web_search: 'Search the web for information using DuckDuckGo. Returns up to 50 results with title, snippet, and URL. Use limit and offset for pagination.'
• Document output schemas. E.g., web_search returns [{title: string, url: string, snippet: string}]. read_file returns {content: string, size_bytes: integer, encoding: string}.
• Add parameter descriptions. E.g., for send_email: 'to' (recipient email address), 'subject' (email subject line, max 200 chars), 'body' (email body, supports plain text and markdown).
• Implement this as a proper MCP server (stdio or HTTP transport) with explicit tool registration via MCP protocol, not a custom agent framework. Current codebase is a LLM agent library, not an MCP server.