Answer out loud before looking back.
One. What are the three things a failing test might mean, and which one is a defect?
Two. Give the test that separates a requirement you can verify from one you cannot.
Three. When is sampling the right call, and what would you do instead with more time?
Four. Why is hand-writing automation now described as the manual step?
Five. What can an agent do right up to the deployment decision, and why not that decision?