The same hackers chained two flaws to reach staff ChatGPT accounts — for a $6,500 bounty

Hacktron's researchers say they chained two previously unknown flaws on July 25 to compromise multiple OpenAI employees' ChatGPT accounts, reaching internal repositories and connected services. They proved the access by using an employee's Codex account to open a harmless pull request in OpenAI's internal monorepo, then stopped testing and reported the flaws.

OpenAI confirmed a fix roughly 14 hours after the report and paid Hacktron a $6,500 bounty on September 1. The September breach was first reported by NBC News — and two months later the firm that pulled it off is on OpenAI's payroll as a partner, not a defendant.