Hackers Used Anthropic's Claude to Breach OpenAI Employee Accounts in Under 72 Hours

A trio of independent security researchers at Hacktron reportedly compromised OpenAI employee accounts in less than 72 hours using Anthropic's Claude Opus 4.8 and 5, the Wall Street Journal reports. The breach gave them access to OpenAI's internal GitHub repository, known as Monorepo, which allegedly houses the company's core algorithmic secrets. Rather than extracting code, the researchers submitted a pull request from a compromised employee Codex account as proof of access. The entry point was Discourse, a third-party platform that hosts OpenAI's community forum. The incident raises fresh concerns about supply-chain vulnerabilities and the use of AI tools to conduct sophisticated cyberattacks.