Safety guardrails blocked Hugging Face's defenders, not the attacker, when an AI agent breached its systems

Hugging Face’s incident response team first turned to frontier AI models to analyze a breach of the company’s production infrastructure, and the models refused to help.