Your question is Detecting Jailbreak Attempts. 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 protocols would you put in place to detect when a patient is trying to jailbreak or push an AI nurse outside of its bounded clinical domain at Amigo?