The Agent Loop

1. What is the correct order of steps in the agent loop?

Decide → Act → Observe → Repeat Act → Observe → Decide → Repeat Observe → Decide → Act → Repeat Repeat → Act → Observe → Decide

2. What is one reason why it is important to have multiple exit conditions in the agent loop?

To ensure the model has more opportunities for error To make sure the model frequently asks for clarification To prevent runaway loops that can't stop To allow for unlimited iterations

3. In the agent loop, which step involves executing the model's choice?

Act Decide Observe Repeat

4. What must happen in the 'Observe' phase of the agent loop?

The model must finalize the goal Current state and context must be gathered The next tool must be decided immediately The loop must exit if it feels incomplete

Continue: Reliability in Agents → Review this lesson Retake quiz