AgentsSeptember 18, 2026via SiliconAngle
Cybersecurity researchers gain access to OpenAI’s GitHub repository using Claude
Why it matters
A real agent exploit—Claude used for multi-step breach automation—demonstrates that agentic AI can execute sophisticated attack chains. This is the kind of silent agent failure (unauthorized access, no alerts) that production teams need to defend against.
Key signals
- Three cybersecurity researchers breached OpenAI's GitHub using Claude
- Repository contained 'OpenAI's algorithmic secrets'
- Files were accessible until June 24, 2026
- Researchers reported findings to OpenAI
- OpenAI patched within 14 hours of notification
- Vulnerability went undetected for months before responsible disclosure
The hook
Claude was used to breach OpenAI's GitHub repo. The repository stayed exposed for months before researchers reported it.
Three cybersecurity researchers used Claude to breach OpenAI Group PBC’s GitHub repository. Sources told the Wall Street Journal today that the repository contains “OpenAI’s algorithmic secrets.” The files were accessible until June 24, the day the researchers reported their findings to the company.…