Two tools with Chinese descriptions and basic schemas. search_feishu_doc has a dynamic enum (good), but both tools lack English descriptions, making them inaccessible to non-Chinese-speaking agents. Parameter descriptions are minimal. Output schemas are not documented in the code. Error handling is absent, no guidance on what to do if a document is not found or the API fails. The tools are READ_ONLY, which is correct, but this is not declared via toolAnnotations. Overall structure is present but falls short of production quality.

• Translate all tool and parameter descriptions to English. Use clear, LLM-optimized descriptions (50 - 200 chars) that explain WHAT the tool does, WHEN to use it, and what it returns. Example: 'search_feishu_doc: Search Feishu API documentation by category. Returns a list of document titles and paths. Use this to discover available APIs before reading detailed docs.'
• Document output schemas explicitly. For search_feishu_doc, specify: returns {documents: [{title: string, path: string}], count: number}. For read_feishu_doc, specify: returns {file_path: string, file_size_bytes: number, warning?: string}.
• Add error handling with recovery guidance. Example: 'If category is invalid, call search_feishu_doc with no args to list available categories.' If a document path fails, suggest 'Try search_feishu_doc to find the correct path.'
• Declare readOnlyHint: true in tool definitions to signal these are safe, idempotent operations. This enables agents to retry without permission concerns.
• Add parameter validation and format hints. For 'path', clarify: 'Document path from search_feishu_doc results (e.g., /docs/api/messaging). Must be a valid path returned by search_feishu_doc.' For 'category', document the enum format more clearly.