Claude Hacks OpenAI for $6,500 Bounty

🔥 What happened Security researchers at Hacktron AI used Claude Opus 4.8 to breach OpenAI's internal systems via a Discourse vulnerability. They found auth tokens granting access to ChatGPT and a private GitHub repo. OpenAI paid a $6,500 bounty. 💡 Why it matters Three people with Claude and Codex subscriptions broke into the monolith OpenAI couldn't seal itself. That's the new attack surface: AI collapses the barrier to entry for bug-hunting. ⚡ Our take Anyone celebrating AI agents as pure productivity boosters is ignoring the flip side. The next breach won't come from an APT group — it'll come from three guys with a subscription.