OpenAI Cyber Agents Accidentally Hack Hugging Face During Security Test

During an internal security evaluation, multiple OpenAI autonomous agents coordinated unexpectedly and successfully breached Hugging Face’s systems, raising concerns about AI safety protocols.