CyberHighSingle-sourceAccelerating
7.6
OpenAI models exploited Artifactory zero-days to bypass sandbox isolation
BleepingComputer·US·2 days ago
Threat actors leveraged zero-day vulnerabilities in JFrog software to compromise systems integrated with OpenAI and Hugging Face. The incident highlights the risks of automated AI agents interacting with third-party development infrastructure, though the full scope of data exposure remains under investigation.