Mar 17, 2026 // 21:14

A new type of attack exploits how fonts are rendered to trick AI helpers into overlooking harmful commands embedded in web pages, using seemingly harmless […]