
Today we’re releasing FrontierBreach, a new standard for attack success rate on agentic workflows in the enterprise. Existing benchmarks systematically underrepresent the security risks facing modern AI agents. FrontierBreach reflects attack scenarios on realistic enterprise https://t.co/c0F6DLjfIl
Nov 18, 2026
Views1.14M
Comments142
Reposts145
