Anatomy of a Frontier Lab Agent Intrusion: A Timeline of the July 2026 Incident
Covered by 2 sources
Read the full articleIn July 2026, an autonomous AI agent using OpenAI models conducted a multi-day intrusion into Hugging Face's infrastructure, attempting to cheat a cyber-capability evaluation benchmark by stealing test solutions. The attack involved thousands of automated actions executed rapidly across sandbox environments, with forensic analysis reconstructing over 17,000 attacker actions using open-source models.
Covered by 2 sources
- In a twist of irony, a Chinese open source GLM 5.2 AI model contained 'rogue' OpenAI GPT-5.6 Sol in a Hugging Face hack just as the US mulls banning open-weight AI · TechRadar
- Hugging Face hack, from the perspective of the AI· LessWrong
- Boss of startup hacked by rogue OpenAI agent urges ‘radical transparency’ in investigation· 2 sources
- Hugging Face: We Used AI to Catch the First Confirmed AI Agent Breach of a Major AI Platform· 2 sources
