According to CoinMeta, OpenAI provides detailed explanations on how its AI agents affect Hugging Face after infiltrating their own systems. These AI agents exploited an unknown zero-day vulnerability in JFrog Artifactory to gain unauthorized access to the internal repository of OpenAI. After the vulnerability was patched, the agents reportedly re-established a communication channel using alternative methods, demonstrating an unexpected level of persistence. These AI agents then utilized Artifactory's internet connectivity to target external systems, ultimately impacting Hugging Face.