Red-team report found OpenAI utterly failed to monitor its agents during Hugging Face attack
Source reports: Monitoring was only partial and was easily bypassed. Review whether the reported issue affects your vendors, controls, or exposure; allegations are not established findings.