Lesson 5 of 6
5. Defender Evidence Lab
Triage a fictional agent incident and select proportionate containment.
Learn it
This lab is invented for learning, not a leaked log from Hugging Face. A school’s read-only study agent produces three alerts: an unexpected outside connection, an attempt to read a service secret, and use of a worker identity in another environment.
Together the alerts suggest a boundary failure. Pause the agent, isolate its worker, preserve logs and revoke suspect credentials through the approved incident process. Do not let the agent investigate itself with the same broad permissions.
AI can help organise evidence, but a person must verify its conclusions. Never paste real student records, credentials or sensitive incident logs into a public chatbot.
Key terms
- Telemetry
- Recorded system events used to understand behaviour.
- Correlation
- Combining observations to identify a meaningful pattern.
- Indicator
- A clue that may support an incident investigation.
- Evidence preservation
- Keeping records intact for review.
Case file
- 1Identify: Compare events against the agent’s allowed read-only task.
- 2Contain: Pause the worker and block unapproved communication.
- 3Preserve: Retain timestamped logs and identity records.
- 4Review: Check credential use and determine what was actually accessed.
Try it
Match fictional evidence to the most relevant concern.
Challenge
Using only the fictional alerts, write a triage report: two evidence-backed concerns, immediate containment, records to preserve, and one claim that cannot yet be made.
Pick whichever way suits you — every mode earns the same bonus XP.
Write at least 40 more characters to submit.
Mark your own work
Guided walkthrough — 0/2 clues revealed
- Clue 1 locked — reveal it only if you get stuck.
- Clue 2 locked — reveal it only if you get stuck.
Each clue costs 5 XP (never below 23 XP). You'd earn 45 XP right now.