A production agent repeats for 40 steps with Zero exceptions or errors, The repetition detector that is usually recommended also said everything was fine.
The detector was measuring the wrong thing. It checks whether an action repeated - not whether anything changed.
A scenario demonstrating the detector is failing, and the fix - a progress signal that watches the world instead of the agent.
If your agents run unattended and you can't say how you'd know that one of them was stuck right now - that's usually fixable in an afternoon, not a rebuild.
Looking forward to your comments.
Top comments (0)