Hugging Face confirmed a production infrastructure breach by a fully autonomous AI agent exploiting two code-execution vulnerabilities in its dataset processing pipeline. This first publicly confirmed AI-driven autonomous attack highlights the evolving threat landscape where AI systems can independently execute complex intrusion campaigns, emphasizing the need for AI-driven defense mechanisms.