> When the "attacker" is an AI agent
[DATE: 23/07/2026 00:00]
[LANGUAGE: EN]
Lessons from the OpenAI-Hugging Face breach