Show Notes
Out of the sandbox and into the fire.
During an OpenAI security evaluation, an AI agent escaped its test environment and breached Hugging Face in search of information that could help it complete its assignment.
No attacker. No malicious instructions. This agent independently concluded that breaking into another company’s systems was the fastest path to success.
Join Matt and David as they dissect the breach step by step and explore what happens when an AI decides security boundaries are merely suggestions rather than rules.
More from State of Cybercrime
ShinyHunters claims it stole 2-3 TB of FBI data and shared a sample with researchers containing information on employees, applicants, and even family members.
If your company runs PTC Windchill, Cl0p may already have your blueprints.
A complete attack chain run by an AI agent – with zero human hands on the keyboard.
The supply chain remains the weakest link. A new automated malware campaign dubbed “Megalodon” poisoned more than 5,500 GitHub repos in just six hours, quietly harvesting cloud credentials, tokens, and sensitive code from environments that implicitly trust what they pull.
275 million users across 9,000 institutions. The #shinyhunters breach of Instructure – the company behind the widely used Canvas LMS – has allegedly exposed 3.6TB of data, making it one of the largest education-sector breaches in history.