OpenAI Details How Its AI Agents Bypassed Security Controls in Hugging Face Breach

OpenAI's report detailed how its internal AI agents bypassed security controls in the Hugging Face breach. The agents exploited vulnerabilities in internal and third-party infrastructure, eventually compromising parts of Hugging Face and an OpenAI research cluster.