Server has 35 tools with complete schemas and descriptions. Naming follows verb_noun pattern consistently (langdock_agent_create, langdock_knowledge_upload_file, etc.). All tools have Pydantic input models with type hints and field descriptions. However, descriptions are often generic (e.g., 'Create a new Langdock AI agent' lacks WHEN/WHY context). Output schemas are not documented, tools return JSON strings without specifying field structure. Error handling is minimal (generic handle_api_error). Tool annotations present but incomplete (missing readOnlyHint on some read tools). Parameters lack actionable constraints in descriptions (e.g., 'JSON array of message objects' should specify expected schema). No pagination guidance for list tools.

• Document output schemas for all tools. Specify which fields are returned (e.g., agent_id, created_at, status) and their types. Include example responses in tool descriptions or a separate schema reference.
• Enhance descriptions with WHEN/WHY context. Replace 'Create a new Langdock AI agent' with 'Create a new AI agent with custom instructions and model. Use this when you need a dedicated agent for a specific task. Returns agent_id for use in subsequent calls.'
• Add explicit parameter constraints in descriptions. For message arrays, specify: 'JSON array of objects with role (user|assistant|system) and content (string) fields. Example: [{"role":"user","content":"Hello"}]'
• Implement pagination for list tools. Add limit (default 20, max 100) and offset parameters. Return total_count and next_offset in responses to enable cursor-based iteration.
• Improve error handling with recovery guidance. Return structured errors: {"error": "user_not_found", "message": "User email not found. Try search_users() with a partial name.", "retryable": false}