OpenAI's cyber-capable models compromised Hugging Face production

How this story unfolded

3 weeks · 75 reports · 69 community posts · 144 of 150 shown

1. Jul 19
   • HuggingFace security incident report: "the attacker was bound by no usage policy, while our own forensic work was blocked by the guardrails"
2. Jul 20
   • World's Largest AI Model Repository Hugging Face Breached by Autonomous AI Agentthehackernews.com
   • Safety guardrails blocked Hugging Face's defenders, not the attacker, when an AI agent breached its systemsventurebeat.com
   • OpenAI says an unnamed long-horizon model tried to break out of its sandbox- and succeeded. During...
3. Jul 21
   • OpenAI Shares Some Alignment Problemsthezvi.substack.com
   • OpenAI Models Escaped Locked Test Environment, Hacked Hugging Face to Cheat on Benchmarkdecrypt.co
   • OpenAI says it accidentally hacked Hugging Face with a new AI systemtheverge.com
   • 9 community posts
4. Jul 22
   • OpenAI's models broke containment and cyberattacked Hugging Face — what enterprises need to knowventurebeat.com
   • OpenAI Says Its AI Models Escaped Sandbox, Targeted Hugging Face to Cheat Benchmarkthehackernews.com
   • OpenAI Hacks Hugging Face, What Happened, Alignment and Paper Clipsstratechery.com
   • OpenAI and Hugging Face Hacking Incident Highlights Growing AI Riskbloomberg.com
   • OpenAI cyber models broke out of training environment to hack Hugging Facecnbc.com
   • GPT-6 Goes Rogue? The HuggingFace Incident, Sans Hypeyoutube.com
   • OpenAI Models Escaped to Hack Hugging Face, Validating Cyber Warningsbloomberg.com
   • OpenAI Models Breach Hugging Face, Sparking Cyber Alarmsbloomberg.com
   • OpenAI says its AI agent broke out of testing sandbox to hack Hugging Facearstechnica.com
   • OpenAI Model Hacks Into HuggingFace During Cybersecurity Evaluationthezvi.substack.com
   • The credential that let OpenAI's agents into Hugging Face exists in most enterprises right nowventurebeat.com
   • 9 community posts
5. Jul 23
   • OpenAI admits AI model hacked Hugging Face, Chinese open-source AI helped investigatetechnode.com
   • AI #178: A Fire Alarm For General Intelligencethezvi.substack.com
   • Oh no...youtube.com
   • OpenAI's Hugging Face hack triggers 'AI Kill Switch' bill in Congresscnbc.com
   • The first known runaway AI agent - or a very bad marketing stunt?simonwillison.net
   • 3 community posts
6. Jul 24
   • Industry Reactions to OpenAI Models Hacking Hugging Face: Feedback Fridaysecurityweek.com
   • AI Can Finally Hack Things by Itselfyoutube.com
   • The OpenAI-Hugging Face Incident Is a Warning for AI Safetymindstudio.ai
   • Why Hugging Face Had to Use a Chinese AI Model to Defend Itselfmindstudio.ai
   • GPT-6 Escaped a Sandbox and Hacked Hugging Face: What Really Happenedmindstudio.ai
   • Was the Model That Hacked Hugging Face Secretly GPT-6?mindstudio.ai
   • OpenAI's Model Escaped Its Sandbox to Hack Hugging Face. Here's Howmindstudio.ai
   • We got Rogue AI Agents hacking HuggingFace and Open-Source models fighting back before GTA 6.
7. Jul 25
   • Why the OpenAI Agent Broke Into Hugging Face: Reward Hacking, Not Malice, Explained for Engineersmarktechpost.com
   • The OpenAI Models That Hacked Hugging Face Were ‘Active on the Internet’ for Dayswired.com
   • 4 community posts
8. Jul 26
   • Hugging Face CEO calls for ‘radical transparency’ after ‘unprecedented’ OpenAI hacktechcrunch.com
   • 2 community posts
9. Jul 27
   • OpenAI's Model Escaped Its Sandbox and Hacked Hugging Face. Here's What Happenedmindstudio.ai
   • OpenAI called the Hugging Face attack unprecedented. But we’ve been here before.technologyreview.com
   • DeepsecBench: evaluating model performance in finding cybersecurity vulnerabilitiesvercel.com
   • 2 community posts
10. Jul 28
    • For Some, So-Called ‘Skynet Day’ Came too Close to Sci-Fi After a Rogue Agent Hacked Into a Startupsecurityweek.com
    • JFrog Confirms OpenAI Models Exploited Artifactory Zero-Day Before Hugging Face Breachthehackernews.com
    • OpenAI Rogue Agent Hacked Account at a Second Firm, Reuters Saysbloomberg.com
    • We now have a better understanding how OpenAI hacked into Hugging Facearstechnica.com
    • Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incidenthuggingface.co
    • 7 community posts
11. Jul 29
    • How independent researchers could investigate AI propensities after misalignment incidentsmetr.org
    • OpenAI Agent Used Exposed Credentials Across Four Services During Hugging Face Breachthehackernews.com
    • OpenAI’s rogue AI agent didn’t stop at hacking Hugging Facetheverge.com
    • Measuring the Tendency of AI Agents to Go Rogueschneier.com
    • Anthropic is finding bugs faster than Microsoft can fix themarstechnica.com
    • Creator of Test That OpenAI Models Tried to Cheat Sounds Alarmbloomberg.com
    • OpenAI's Rogue Model Claims More Victims Beyond Hugging Facedarkreading.com
    • 3 community posts
12. Jul 30
    • Highlights From The Discourse On The Hugging Face Incidentastralcodexten.com
    • The AI Safety Rule That Left Hugging Face Defenselessmindstudio.ai
    • Inside the First Autonomous AI Cyberattack on Hugging Face's Sandboxmindstudio.ai
    • 2 community posts
13. Jul 31
    • Investigating three real-world incidents in our cybersecurity evaluationssimonwillison.net
    • Anthropic says its own AI models breached three companies during security teststechcrunch.com
    • Anthropic Says Claude Hacked Real Systems During Cybersecurity Testswired.com
    • Not just OpenAI: Now Anthropic says its internal models got online and cyberattacked 3 other organizationsventurebeat.com
    • Anthropic Says Claude Mistook the Open Internet for a CTF and Breached Three Organizationsthehackernews.com
    • AI #179 Part 2: Hearing The Fire Alarmthezvi.substack.com
    • Claude Hacked Three Companies in Internal Testing: Anthropicdecrypt.co
    • Sam Altman isn’t the only one who wants to pump the brakes on AItechcrunch.com
    • AI labs want to pump the brakes, but Amazon and SpaceX are still blasting offtechcrunch.com
    • 3 community posts
14. Aug 1
    • Nobody Knows if OpenAI’s and Anthropic’s AI Hacking Sprees Are Illegalwired.com
    • 2 community posts
15. Aug 2
    • July 2026 newslettersimonwillison.net
    • Further Developments About Internal AI Models Hacking Thingsthezvi.substack.com
    • AI firms must answer for rogue bots, says boss of hacked company
16. Aug 3
    • Here’s why AI agents lie and cheat to reach their goalstechnologyreview.com
    • The OpenAI Hack Shows the Genie Is Out of the Bottleschneier.com
    • More on the OpenAI Agent’s Attack on Hugging Faceschneier.com
    • 5 community posts
17. Aug 4
    • OpenAI Says Models Breached Boundaries During Outside Testingbloomberg.com
    • 2 community posts
18. Aug 5
    • Claude Mythos 5 Tried to Backdoor a Real Open-Source Project in Testing, Then Vouched for Itselfthehackernews.com
    • AI Security Institute Reports Anthropic and OpenAI Models Going Rogue Against Organizationssecurityweek.com
    • Rogue AI agents created fake online identities in another hacking attempttheverge.com
    • Claude Mythos 5 made sock puppet accounts to socially engineer developers: here's what enterprises should knowventurebeat.com
    • Anthropic’s AI used fake identities, malware in rogue attack on GitHub projectarstechnica.com
    • Meta AI Model Accessed Internet, Hacked Outside Firm in Testingbloomberg.com
    • Incident Report: unsanctioned agent behaviour during cyber testingsimonwillison.net
    • 6 community posts
19. Aug 6
    • Third-party cyber evaluations involving OpenAI modelssimonwillison.net
    • An AI model from Meta also hacked another company during testingsimonwillison.net
    • OpenAI Models Joined Forces Months Ahead of Hugging Face Hackbloomberg.com
    • 6 community posts