RESEARCHERS USED CLAUDE TO HACK OPENAI

Summary from the TLDR InfoSec newsletter

Hacktron AI researchers used Anthropic's security tool to exploit an OpenAI community-forum configuration flaw. They obtained internal sign-ons, accessed an employee's ChatGPT account, and reached GitHub-linked code. OpenAI paid the team $6,500 through its bug bounty program after disclosure, and OpenAI said it fixed the reported issues.