Hacktron AI Used Claude to Expose OpenAI Account and Repository Flaws

Hacktron AI researchers used Anthropic's Claude and OpenAI's own GPT-5.6 Sol to expose a chain of flaws that let them take over OpenAI employee ChatGPT and Codex accounts and reach an internal repository, before reporting the issue through OpenAI's bug bounty process. The company fixed the OpenAI-side issue the same day it was reported and later paid a $6,500 bounty.