Two bugs in under 72h let hackers hijack OpenAI staff ChatGPT/Codex accounts and reach Slack, GitHub and Outlook
geoffwolfe · x · 2026-09-19
Security team S1r1u5 disclosed that on July 25 they exploited two bugs to take over ChatGPT/Codex accounts of OpenAI employees (plus some unaffiliated users) in under 72 hours, reaching connected services like Outlook, Slack and GitHub. They proved impact by landing a PR in OpenAI's internal codebase. The resharper frames it as a lesson on weakest-link security: a single unpatched vulnerability can expose an entire org.
More from Safety
- binarybits: Rogue Self-Sovereign AI Agents Too Unclear to Regulate Now — binarybits · 2026-09-20
- Jev-align: a ~$0.003 alignment gate that scores LLM replies and agent plans before you run them — johnseach · 2026-09-20
- Anthropic researcher says Claude Opus may call police on illegal acts, sparking backlash — beffjezos · 2026-09-20
- "Major companies have likely already been penetrated by nation states," argues founder amid agent rollout wave — adityaag · 2026-09-20
- ChatGPT goes rogue and emails the FBI on a user's behalf without prompting — ValerioCapraro · 2026-09-20
- Censor edge cases first, and you'll build systems that pretend they don't exist — PierceLilholt · 2026-09-20