Skip to content
AI Security

Interview question · Prompt Injection & Jailbreaks

How do capability-based designs such as CaMeL bound what a prompt injection can do?

Review capability metadata, trusted labels, derived values and actual release policy. Work through allowed sharing, forbidden audiences, branch influence and uncovered runtime paths.

expertNewUpdated Nov 2026

Review capability metadata, trusted labels, derived values and actual release policy. Work through allowed sharing, forbidden audiences, branch influence and uncovered runtime paths.

Make your first attempt

What would you say before reading the answer?

  1. 01 State the assumptions your answer depends on.
  2. 02 Explain the mechanism and choose a control.
  3. 03 Name a condition that would change your decision.

The worked answer includes an approach, scoring criteria and follow-up questions. Compare the reasoning with your own.

10 answers per topic instead of 5, plus saved progress and bookmarks · costs nothingor unlock all 529 remaining answers · $25

Every worked answer in all 17 topics, every lesson in every course, the full concept curriculum, the complete practice-test bank, and the downloadable security field guides.

6 months · One payment · No auto-renewal

What if · the follow-ups in this answer

The answer works through 4 follow-ups like an interviewer would

Tessa prepares a fictional project briefing from a supplier description and a staff-only note. Internal display is permitted under the current grant; external sharing requires an authorized content-release decision.

  1. Tessa types the supplier's address herself. May the staff-only note now be sent there?
  2. A confidential Boolean controls whether a constant message is sent. Does checking only the message body cover the flow?
  3. The interpreter checks sends, but an export helper calls the transport directly. Is the intended release rule enforced everywhere?
  4. What would your assurance statement include?

Each rung has a nudge, a model answer, the principle that carries over and the weak answer to avoid. Premium opens every answer and ladder.

UP NEXT ON YOUR JOURNEY
Next in this trackTrace how instructions can reach a model through emails, web pages and tool outputs in a real application.Next in this trackDesign an email assistant that bounds the harm from prompt injection.Next in this trackHow would you test whether a RAG assistant is vulnerable to indirect prompt injection, without harming production?
DISCUSSION · 0

No comments yet. Share a correction or a different way to answer it.