Preventing Infinite Loops in AI Agents: Fail-Safe Solutions Needed

I’ve noticed that Lindy’s have a really high chance of entering into infinite loops if they touch other AI agents. Are there any way to build in “fail safe’s” into our models so that if it exceeds a certain number of runs or credits then it stops. Marvin A. I have a lindy that hallucinated adding another lindy of mine. Now they are going back thanking each other and I’m not able to stop it 😞