Module 07 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. When a crew produces a poor result, the first thing to inspect is…
- A. The final output
- B. Each task's individual output, to find where quality first dropped
- C. The model
- D. The cost
2. Per-agent token and time accounting reveals…
- A. Nothing useful
- B. Which role is driving cost and latency, so optimisation is targeted
- C. Answer quality
- D. Tool correctness
3. Logging tool calls with their arguments matters because…
- A. It is required
- B. Wrong arguments are a common failure that is invisible in the final output
- C. It reduces cost
- D. It speeds runs
4. Run inspection data typically contains…
- A. Only metrics
- B. Business content and possibly client data, so it needs access control
- C. Nothing sensitive
- D. Only timings
5. The clearest early warning of a looping crew is…
- A. High latency alone
- B. Repeated near-identical task outputs or tool calls
- C. Large prompts
- D. Many agents
---
Answer Key and Explanations
Check these only after attempting every question.
1. B - Each task's individual output, to find where quality first dropped
In a chain, the visible failure is usually downstream of the actual one.
2. B - Which role is driving cost and latency, so optimisation is targeted
Without attribution, optimisation is guesswork.
3. B - Wrong arguments are a common failure that is invisible in the final output
A tool called correctly with wrong inputs produces confidently wrong downstream work.
4. B - Business content and possibly client data, so it needs access control
Automation logs quietly accumulate the very material the automation handles.
5. B - Repeated near-identical task outputs or tool calls
Repetition without progress is the signature of a non-converging delegation cycle.
Practical Check
Apply this module to your own work: complete the module activity for *Observability and Run Inspection*, 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.