Threatline

AI Browsers Vulnerable to Zero-Click Agent Hijacking

Attackers can take control of agents through malicious instructions hidden in content supplied to AI browsers, and there's no simple fix for the threat.