Skip to content
AI Security

Interview question · Agent, Tool & MCP Security

How do you test a function-calling agent for unintended tool use, and what number do you report?

Define current task authority and forbidden effects, test policy and agent behavior together, and report proposals, outcomes and useful completion with honest denominators and uncertainty.

mediumNewUpdated Nov 2026

Define current task authority and forbidden effects, test policy and agent behavior together, and report proposals, outcomes and useful completion with honest denominators and uncertainty.

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

A synthetic 120-task evaluation records 18 tasks with forbidden proposals, six with forbidden dispatch and four with confirmed forbidden effects. Two dispatched outcomes remain unresolved.

  1. What is the confirmed forbidden-effect rate per task?
  2. Would zero failures in a different 40-run search prove low production risk?
  3. Should a model be rewarded for refusing all 120 tasks?
  4. What evidence belongs in the release decision?

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 trackHow does trust propagate in a multi-agent system, and where do you put the boundaries?Next in this trackHow should untrusted documents and web content enter an agent workflow?Next in this trackWhat are the security risks of agents that operate a browser or a desktop, and how do you contain them?
DISCUSSION · 0

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