S4e22 Huggingface Compromised By Agentic Attack Gold Eagle Program

Summary

00:00 The Adversarial Podcast 00:58 Hugging Face's AI-driven incident disclosure 03:48 What makes an Leading AI companies Anthropic, OpenAI, and xAI recently called for a slowdown in the industry. This follows an incident in July ... OpenAI & Hugging Face security incident explAIned Watching this back - I misspoke when I called sandboxes “secret” lol I meant ... OpenAI is imposing stricter controls in response to the Hugging Face rogue agent incident. On Jeremy Kahn ... In July, Hugging Face was breached. On July 16 it published a post saying it had been In July 2026, a swarm of OpenAI's AI agents escaped a cybersecurity test and broke into Hugging Face. No human directed it. This could get scary really quickly. A full week went by before OpenAI realized its own agents were This is a video version of a post I wrote. If you want to read the original you can check it out here: ... A swarm of 700 OpenAI agents escaped their sandbox and breached Hugging Face using nearly a million chained links. New details about the Hugging Face 1200 AI agents inside OpenAI's infrastructure were supposed to be acting in isolation. They were not. They discovered each ... An AI agent was given an impossible hacking challenge. Instead of giving up, it left itself a note in shared storage — and other ... In July 2026, about 700 OpenAI agents took part in an