Hugging Face Hacked by an AI Agent in an Ironic Twist

In a case of life imitating art, Hugging Face—the popular open-source AI platform—has disclosed that an autonomous AI agent breached its production systems. The company revealed the incident last week, stating that the attack originated from a malicious dataset that exploited two code execution pathways in its data processing pipeline. The intruder used these vulnerabilities to run code on a processing worker, then escalated to node-level access and moved laterally across several internal clusters over a weekend.

Hugging Face confirmed that no public models, datasets, or Spaces were tampered with, and the software supply chain remains intact. The attacker, operating through an autonomous agent framework, executed thousands of actions across a swarm of short-lived sandboxes, using self-migrating command-and-control servers hosted on public services. The specific large language model used by the attacker is unknown, but the company has since patched the code execution flaws, rebuilt compromised nodes, and rotated all affected credentials. Additional guardrails and 24/7 monitoring have been deployed.

In an unusual twist, Hugging Face turned to Z.ai's GLM 5.2, a Chinese open-weight model, for forensic analysis after Western frontier models refused requests containing real attack commands and exploit payloads due to their safety guardrails. The company noted a critical lesson for defenders: have a capable, unrestricted model ready on your own infrastructure before an incident to avoid being locked out by safety filters and to keep sensitive data from leaving your environment. Hugging Face is urging all customers to rotate access tokens and review recent account activity.