Hacktron researchers used Claude Opus 5 to breach OpenAI staff accounts via chained flaws
Three researchers from security firm Hacktron used Anthropic's Claude Opus 5 to chain two flaws — a libheif image parsing bug in OpenAI's public Discourse forum and a shared SSO login weakness — to take over ChatGPT and Codex accounts of OpenAI employees and access an internal code repository in under 72 hours. They reported the vulnerabilities responsibly, submitted a harmless pull request as…
Key points
- Hacktron researchers used Claude Opus 5 to chain a forum image bug and SSO flaw to access OpenAI staff accounts and internal code in under 72 hours.
- OpenAI confirmed a fix within 14 hours of the report and paid a $6,500 bounty on September 1 for the responsibly disclosed vulnerability.
- Digital Trends reports the operation cost ~$3,000 in Anthropic AI usage; The Rundown notes it happened in July and links it to OpenAI models allegedly hacking Hugging Face.
While all sources agree on the core event, Digital Trends adds that the researchers spent approximately $3,000 on Anthropic's systems to carry out the operation and includes criticism from Hacktron's Fabian Faessler that OpenAI's CISO initially called them unprofessional before apologizing. The Rundown notes the breach occurred in July and references a parallel incident where OpenAI's own models allegedly hacked Hugging Face, though this claim is not echoed in the other articles.
The story so far
3 episodes →- Hacktron researchers used Claude Opus 5 to breach OpenAI staff accounts via chained flawsthis story
Claude Opus 5 Helped Researchers Take Over OpenAI Staff Accounts via Chained Flaws
thehackernews.com · 19 September 2026
Loading the full article…
This text was published by thehackernews.com and written by The Hacker News. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
3sources- OpenAI goes from hacker to hackedNewsletter · The Rundown AI ·
- Experts used AI to break into OpenAI. They say it was cheap and a red alarm for the industryPress · digitaltrends.com ·
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Enterprise & Industry
All →- Cyara’s CEO outlines AI agent oversight as a workforce model · 1 src
- Arcana adds institutional portfolio data to Claude’s connectors · 1 src
- Anthropic brings Claude inference to India via Amazon Bedrock · 6 src
- Infor adds industry-specific AI agents to Velocity Suite, citing generic AI gaps · 1 src
- Tech Mahindra trains 12,500+ staff on Google’s Gemini Enterprise agents · 1 src
Comments
via GitHub Discussions