Classic threat modeling assumes a system that holds still. Autonomous agents don't — and the July 2026 Hugging Face incident shows why design-time and runtime have to be one continuous picture.
