Hackers used Anthropic's Claude to breach OpenAI's source code for $6,500 reward

Three hackers breached OpenAI's source code repository on GitHub using Anthropic's Claude AI, earning a $6,500 reward through OpenAI's bug bounty program. The hackers accessed OpenAI's "Monorepo," described as the company's "secret sauce" for faster model performance. Hacktron AI's Mohan Pedhapati said the team used Claude and Codex subscriptions, according to the Wall Street Journal. The incident adds to a series of security breaches involving OpenAI and Anthropic, both of which have advocated for safety guardrails in AI development. The companies warn AI models are finding cybersecurity vulnerabilities faster than they can be patched.