OpenAI Tools That Compromised Hugging Face Remained Online for Days

Security researchers found that certain OpenAI tools were used to breach Hugging Face. The compromised tools were observed operating on the internet for several days. The intrusion gave unauthorized access to code repositories on the platform. Hugging Face detected the activity after noticing unusual traffic patterns. OpenAI confirmed that its tools had been compromised in the incident. The breach raises concerns about supply‑chain security in AI ecosystems. Both companies are conducting investigations to determine the root cause. Users are advised to monitor their accounts for any suspicious behavior.