Both tools have minimal descriptions (under 30 chars) and trivial implementations that return random Spanish phrases regardless of input. Schemas are present but parameters lack depth. No error handling, no output documentation, and tool behavior does not match descriptions. This is a toy/demo server, not production-ready.

• Expand tool descriptions to 50 - 150 characters. Example: 'ask_question: Ask Manuel a question and receive a response. Use this to gather information or clarification. Returns a string response from Manuel.'
• Document the output schema explicitly. Add a return type annotation or schema field indicating the tool returns a string with the response text.
• Implement actual tool logic that processes input meaningfully. Currently both tools ignore the question/task parameter and return random phrases. This defeats the purpose of having parameters.
• Add parameter descriptions explaining expected format and constraints. Example: 'question: A clear, specific question to ask Manuel (1 - 500 characters, English or Spanish).'
• Implement error handling for edge cases (empty input, oversized input, invalid characters). Return structured error responses with actionable guidance.