Why it matters

An agent can drift from its intended role through compromise, misalignment or poisoned context while each individual action still looks legitimate.

How to prevent it

  • Define each agent’s scope and enforce it with least-privilege permissions and policy outside the model.
  • Monitor behavior against a baseline and alert on actions outside the agent’s role.
  • Keep the ability to suspend an agent and revoke its credentials quickly.

Guides on this site

Main subject · Guide

Contain cascading failures and rogue agents

Stop a looping, compromised or drifting agent from spreading harm: kill switches, budget and rate limits, circuit breakers, behavior monitoring, blast-radius limits, quarantine, rollback and incident response, all enforced outside the agent.

References

  1. OWASP Top 10 for Agentic Applications for 2026 (OWASP GenAI Security Project, Agentic Security Initiative)

List data from the publisher. A guide is listed as the main subject when the weakness or risk is what the guide is about, and as covering it when a dedicated section or worked example addresses it.