TechResearchers used Anthropic's Claude Opus 5 to breach OpenAI's systems in 72 hours
Hacktron AI researchers chained a libheif vulnerability with sign-on misconfiguration to access OpenAI employee accounts and internal GitHub. Claude Opus 5 rapidly produced working exploit. OpenAI paid $6,500 bounty and credited researchers.
TLDR
Dramatically illustrates how AI coding agents accelerate vulnerability discovery and exploitation, with model version upgrades significantly improving exploit speed. Highlights competitive dynamics between AI labs and validates bug bounty programs as mechanisms for identifying security gaps before malicious exploitation.
Combined views
18.3K
2 Sources, first seen 19h ago
Researchers used Anthropic's Claude Opus 5 to breach OpenAI's systems in 72 hours
Hacktron AI researchers chained a libheif vulnerability with sign-on misconfiguration to access OpenAI employee accounts and internal GitHub. Claude Opus 5 rapidly produced working exploit. OpenAI paid $6,500 bounty and credited researchers.