ModelsSep 18, 2026 · 12:508 sources

Three Researchers Breach OpenAI Monorepo Using Claude Opus 5

From huggingnews.com

Image: huggingnews.com
AI;DR

Independent security experts from Hacktron AI used Anthropic's Claude Opus 5 to breach OpenAI's private GitHub monorepo on July 25, chaining a Discourse flaw and an image upload bug to acquire authentication tokens. The attack cost less than $3,000 in AI tokens, and OpenAI paid a $6,500 bounty after confirming only limited reads of private repository metadata and code changes with no model weights exposed. The incident was part of a broader investigation named HEIF Heist into the libheif library, which the researchers claim also enabled them to hack Meta, Slack, and GitHub.

Read the sourceSign in to suggest

Sources

+ 4 more
  1. SUPPORT@mtslive“Three researchers at Hacktron AI, operating under OpenAI’s bug-bounty safe harbor, used Claude Opus 5 in late July”x.com
  2. SUPPORT@rohanpaul_ai“OpenAI said its review found only “limited reads” of private-repository metadata and code changes; no model weights were believed exposed”x.com
  3. SUPPORT@yuchenj_uw“Their entire hacking cost less than $3000 in tokens”x.com
  4. SUPPORT@tradfi“OPENAI PAID SECURITY RESEARCHERS A $6,500 BOUNTY FOR UNCOVERING VULNERABILITIES”x.com

One ranked edition, every morning

The stories that matter today. Free, cancel anytime.

Telegram @aidr_today