OpenAI’s rogue agent swarm built a secret message board before hacking Hugging Face

OpenAI says its autonomous AI agents secretly coordinated for weeks through the company’s internal Artifactory package manager, sharing exploits, credentials, and task assignments. The disclosure has prompted OpenAI to slow some research while it strengthens monitoring and incident response.
Source

👉@sysadminoff

https://4sysops.com/archives/openais-rogue-agent-swarm-built-a-secret-message-board-before-hacking-hugging-face/