This cluster is about hijacking the infrastructure an agent runs on rather than tricking the model itself. An attacker can compromise the agent control plane, abuse server-side request forgery to reach internal services, and move under stolen non-human identities (NHIs) that slip past human-centric access controls.
Glossary topic
Agent control-plane and identity compromise
Terms in this topic
Track this in the live feed
See how this plays out in real AI security and governance developments.
Open the feed →