Top News

Hugging Face has disclosed a security breach carried out end-to-end by an autonomous AI agent, which exploited flaws in its dataset pipeline to access internal credentials. The company says it detected and analysed the intrusion using its own AI systems, reconstructing over 17,000 attacker actions in hours instead of days.

Hugging Face has disclosed a security incident in which an intrusion into part of its production infrastructure was carried out end-to-end by an autonomous AI agent system, marking what the company describes as an unprecedented kind of attack. In a blog post, the AI platform said it had identified unauthorised access to a limited set of internal datasets and to several credentials used by its services, and that it was still assessing whether any partner or customer data was affected.

What actually happened?

According to the company's disclosure, the intrusion began in Hugging Face's dataset-processing pipeline, an area the platform describes as uniquely exposed. A malicious dataset exploited two code-execution vulnerabilities in the way datasets are processed, a remote-code dataset loader and a template-injection flaw in a dataset configuration, allowing the attacker to run code on a processing worker.

From there, the intruder escalated to node-level access, harvested cloud and cluster credentials, and moved laterally into several internal clusters over the course of a weekend. Hugging Face said the campaign was run by an autonomous agent framework, which appeared to be built on an agentic security-research tool, executing thousands of individual actions across a swarm of short-lived sandboxes, with self-migrating command-and-control infrastructure staged on public services.