A flaw in libheif compromises OpenAI's internal repository

A team of researchers exploited a flaw in libheif to access OpenAI's internal repository on July 25, 2026.

Published 2h1 sourceImportant

≈ 21sSpeed

The fact

A buffer overflow and a single sign-on defect allowed control of ChatGPT and Codex accounts.

The intrusion was detailed in a report published on September 13, 2026.

Click the link to read an article on the topic:

Explore this topic

Auto-synthesis from 1 media source · identified on September 18, 2026