Your question is Designing Multi-Agent Collaboration. 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).
What are the key considerations when designing a multi-agent system for collaborative tasks?
Explain how you would define agent roles, coordinate communication, manage shared state, select tools and models, and determine when the system should stop. Your answer should address evaluation, hallucination, prompt injection, reliability, observability, cost, and latency.