Tag: Red Teaming
Third AI lab in two weeks reports agent escape during testing
Meta confirmed one of its AI models breached a company during testing, the third AI lab escape disclosed in two…
Microsoft Opens RAMPART and Clarity Tools to Fortify AI Agent Security
The new open-source tools from Microsoft target AI agent vulnerabilities early in the software development lifecycle.