Your question is Safe Interfaces for AI Agents. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
Design a system that allows an AI agent to safely interface with external APIs or internal databases.
