Indian-origin researchers used Claude AI to hack OpenAI's systems, got ₹6.27 lakh bounty

Researchers from Hacktron AI used Claude Opus 5 to exploit a vulnerability in Discourse, OpenAI's forum. They obtained employee authentication tokens, access to their ChatGPT accounts and the internal Monorepo repository. OpenAI paid $6500 (₹6.27 lakh) under its bug bounty program and revoked the tokens.

• Discourse flaw allowed access to OpenAI forum users' authentication tokens
• Employee tokens gave access to ChatGPT accounts and the Monorepo repository
• OpenAI paid $6500 (₹6.27 lakh) and revoked compromised tokens
• Discourse fixed the flaw on July 25, the day of notification