The purpose is to find things. A mock that finds nothing has failed, not succeeded.
Say that to the candidate at the start. A candidate who thinks the goal is to get through cleanly will hedge, and hedging hides exactly what this lesson is for. A candidate who understands the goal is to surface problems while they are still cheap will engage differently, and the lesson is worth three times as much.
The most diagnostic single move is the level change. Standard 3. Candidates who have only ever taught a Simulated Student played by their own instructor have usually converged on one pitch. An examiner will not warn them before changing Student, and neither should you.
Count the self-critique. Standard 7 is a number, and it is worth recording as a number on the deficiency list: found six of nine. Over a program that number is the best single predictor of whether a new instructor will self-correct without supervision, which is what post-certificate mentorship inherits.
Write the deficiency list during the lesson, not afterwards. Per standards/CRITIQUE-STANDARD.md an observation stated as behavior is a fact, and the same observation reconstructed two hours later is an impression. The candidate is entitled to the fact.
You do not set Tracker Status and you do not write the recommendation. Both are the Supervising Instructor's, named and dated, per cross-cutting requirement 9. Your output is the deficiency list and one labelled sentence of opinion. Lesson 42 is where the decision is made.
Hand the deficiency list to the Supervising Instructor and to the candidate the same day, and file it. It survives certification, and post-certificate mentorship consumes it in another project.