Design a runtime policy engine for action approval, refusal, and escalation.

Instruction: Describe how you would decide at runtime whether the assistant may act, refuse, or hand off.

Context: Assesses whether the candidate can design a practical architecture and explain the main tradeoffs. Describe how you would decide at runtime whether the assistant may act, refuse, or hand off.

Official answer available

Preview the opening of the answer, then unlock the full walkthrough.

I would make the runtime policy engine explicit and inspectable. The system should be able to explain why an action...

Upgrade to view official answer

Related Questions