Open cardosec

Case B425B9 · AI security · L1 Foundations

Where an LLM App Is Vulnerable

Practise as: Explain it · InterviewLevel: L1 Foundations

Interview questionWalk me through how a typical LLM-powered app works and where its security boundaries are.

  1. 01 What is it?
  2. 02 How is it abused?
  3. 03 How do you stop it?
Practise this card As Interview
What a strong answer covers

Try it out loud first. Then check yourself:

  1. The app sends the model one block of text: system prompt, conversation, retrieved documents and tool results together.
  2. The model cannot reliably tell instructions from data in that block, which is why prompt injection works at all.
  3. So the app, not the model, must enforce access: check permissions in code before retrieving data or running a tool.
  4. Treat model output like user input: validate it before it reaches a browser, a database, a shell or another system.

If the interviewer pushes back

  • A product manager wants the model to decide which records a user may see. What do you say?

Go deeper

More AI security cards

All AI security cards →

cardosec draws a security topic and gives you a clock: explain it out loud with no notes, then see what you covered and what you missed. Free during early access.

Reviewed 2026-10-09