Hacktron reported that on July 25, 2026 it chained two critical vulnerabilities to take over multiple OpenAI employees' ChatGPT accounts and open a proof-of-concept pull request in an internal OpenAI repository. The chain began with a libheif heap buffer overflow in OpenAI's Discourse forum, found and exploited using Claude Opus 4.8 and Opus 5. OpenAI fixed its side within about 14 hours and paid a $6,500 bounty.
The root cause is a supply-chain gap: an upstream fix was not labelled as security-relevant, so Debian 12 images shipped vulnerable libheif 1.19.7. Services processing user-supplied HEIC or HEIF images should check libheif. The excerpt does not detail the SSO flaw itself.