Single tool 'query_user' has a basic schema and description but significant gaps. Description is in Chinese and lacks English clarity for international LLM use. Input schema uses Zod but is not properly documented in the output. No output schema documented. Error handling is minimal (returns text in content array but no structured error classification). Tool name lacks action verb clarity ('query' is weak; 'get_user' would be stronger). No parameter constraints (userId accepts any string; should validate against enum ['001', '002', '003']). No pagination, no idempotency hints, no permission gates. Resource registration present but not a tool.

• Rename 'query_user' to 'get_user' to follow verb_noun convention and improve LLM intent recognition.
• Translate all descriptions and error messages to English. Use: 'Retrieve user information by ID. Returns name, email, and role. Valid IDs: 001, 002, 003.'
• Add enum constraint to userId parameter: {type: 'string', enum: ['001', '002', '003'], description: 'User ID (001, 002, or 003)'}
• Document output schema explicitly: {type: 'object', properties: {id: {type: 'string'}, name: {type: 'string'}, email: {type: 'string'}, role: {type: 'string'}}}
• Return structured error responses with recovery hints: {error: 'user_not_found', message: 'User 999 not found. Valid IDs: 001, 002, 003', valid_ids: ['001', '002', '003']}