Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident

A lot of people have jumped to the illogical conclusion that all this is a reason to ban open-source models. I think that's silly. When OpenAI's agents broke into Hugging Face, the models Hugging Face reached for first, Claude Opus and Fable, refused much of the work because of their guardrails. It was GLM-5.2, an open-weight Chinese model running on Hugging Face's own machines, that helped them decode the attack. B.