This is a Tauri desktop application, not an MCP server. The codebase implements a UI tool for managing MCP servers via Tauri's invoke() API, which is a local IPC mechanism, not the Model Context Protocol. The six tools (get_mcp_servers, set_server_enabled, get_desktop_mcp_servers, set_desktop_server_enabled, get_projects, set_project_disabled_servers) are Tauri commands, not MCP tools. While the tool definitions are present in src/lib/api.ts as TypeScript function signatures, they lack MCP-compliant JSON Schema definitions, parameter type annotations, and proper tool registration. The descriptions are partially translated from Japanese and provide operational context but fail to meet LLM-optimization baselines. Input schemas are inferred from TypeScript interfaces, not explicitly defined. Most critically, this is not an MCP server, it is a desktop application that manages MCP server configurations. Evaluating it against MCP standards is architecturally mismatched.

• This project must first be converted from a Tauri desktop app to an actual MCP server. Implement transport support (stdio, HTTP, or SSE) using an MCP SDK (e.g. mcp-rust, mcp-typescript) or a raw MCP protocol library.
• Define all six tools with explicit JSON Schema input and output schemas. For example, set_server_enabled should declare: input {type: 'object', properties: {name: {type: 'string', description: 'Server name (must exist in get_mcp_servers result)'}, enabled: {type: 'boolean', description: 'true to enable, false to disable'}}, required: ['name', 'enabled']}
• Translate all descriptions to English and optimize for LLM comprehension. Expand descriptions to 50 - 150 characters, stating action, use case, and dependencies. Example: 'Retrieve all registered global MCP servers. Returns both enabled and disabled servers. Call this first to discover available servers or verify a server name before calling set_server_enabled.'
• Document output schemas for all tools. Example for get_mcp_servers: 'Returns array of {name: string, enabled: boolean, command?: string, args?: string[]}. Use the name field to identify servers for set_server_enabled calls.'
• Add error handling with recovery guidance. For set_server_enabled, document: 'If the server name is not found, call get_mcp_servers to retrieve valid names. If permission denied, the user may lack write access to the config file.'
• Implement permission checks and audit logging. Verify the calling user/agent has authority before modifying server settings. Log who changed what, when, and why for compliance.