Module 06 Knowledge Check
5 questions. Pass mark 4 out of 5. Answer every question before checking the answer key below, then retry after reading the feedback.
1. The gate belongs…
- A. at the end of the run, with the summary
- B. immediately before the first irreversible step
- C. at the start, before any work
- D. after each step
2. The review-card field most often missing is…
- A. the timestamp
- B. what happens if the reviewer does nothing
- C. the agent's reasoning
- D. the model version
3. The agent's reasoning should…
- A. lead the review card
- B. stay in the trace - it is long, persuasive, and not evidence
- C. be summarised for the reviewer
- D. replace the evidence list
4. A rejection is actionable when it carries…
- A. a polite tone
- B. what was wrong, the correct value, and the next action
- C. the full trace
- D. the reviewer's name
5. Edit rate and time-on-screen falling towards zero means…
- A. the system has improved
- B. approvals have become reflexive and the gate has stopped being a control
- C. reviewers are more experienced
- D. the review card is clearer
---
Answer Key and Explanations
Check these only after attempting every question.
1. B - immediately before the first irreversible step
A review after the notice was sent is a notification, however much it looks like a control.
2. B - what happens if the reviewer does nothing
A reviewer who does not know the default approves to make the decision go away.
3. B - stay in the trace - it is long, persuasive, and not evidence
Reasoning tells the reviewer what the agent thought rather than what is true.
4. B - what was wrong, the correct value, and the next action
Without all three the retry has the same information as the first attempt and produces the same output at full cost.
5. B - approvals have become reflexive and the gate has stopped being a control
The cause is the request rate, so the fix is fewer gates rather than more diligence.
Practical Check
Apply this module to your own work: complete the module activity for *Human Approval and Escalation*, then write one sentence naming what your result shows and one naming what it does not.
Strong Answer Pattern
A strong answer names the task, the evidence used, the check performed, and the remaining limitation. It avoids "proved", "guaranteed", or "always" unless the evidence genuinely supports it.