Hacktron researchers used Claude Opus 5 to exploit vulnerabilities, gaining unauthorized access to OpenAI staff accounts through chained flaws.
- Researchers at Hacktron leveraged Claude Opus 5 to exploit vulnerabilities in OpenAI’s systems.
- The attack involved a series of chained flaws that allowed unauthorized access to staff accounts.
- This incident highlights significant security concerns surrounding AI tools and their potential misuse.
