Attacks that use crafted content to alter an AI system's intended behavior
Web Security
+7
Jul 29, 2026
•
6 min read
Browser agents need a security boundary between what they read and what they are allowed to do
