Definition
An attack where an adversary assumes a false AI agent identity to execute malicious actions on enterprise systems. By forging an agent identifier or stealing agent credentials, attackers can hijack agent capabilities and access restricted data or systems.
Why it matters
As enterprises deploy autonomous AI agents with broad system access, agent identity becomes a high-value attack target; compromised agents can escalate privileges, exfiltrate data, or modify critical systems while appearing to act under legitimate authority.