Skip to content
AI Security

Interview question · Agent, Tool & MCP Security

What is scope drift in an autonomous agent, and how would you detect it from the agent's traces?

Compare actions with current grants and policy, distinguish unusual behavior from violations, and handle legitimate task revisions. Verify observation coverage and respond to the affected operation.

mediumNewUpdated Nov 2026

Compare actions with current grants and policy, distinguish unusual behavior from violations, and handle legitimate task revisions. Verify observation coverage and respond to the affected operation.

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

Grant G-1 permits drafting a project summary. The owner later authorizes publication of the approved report to a named team workspace through controlled grant G-2.

  1. Is publication under G-2 necessarily scope drift?
  2. Can a retrieved note create the same authority?
  3. Should an unusually long drafting run be labeled malicious drift?
  4. What should the monitor demonstrate before release?

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 trackBeyond a single injected prompt, how do attackers corrupt AI agents through memory, tools, retrieved knowledge and other agents?Next in this trackHow do AGENTS.md, rules files and editor or MCP config in a repository become an attack path against coding agents?Next in this trackWhat does the same-origin policy protect when an AI agent acts across logged-in sites?
DISCUSSION · 0

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