Critical Exploit Chain Compromises OpenAI Employee Accounts and Repositories
September 17, 2026
Attackers chained vulnerabilities in libheif, ImageMagick, and Discourse SSO to compromise OpenAI employee accounts. This enabled unauthorized access to internal repositories, including the openai/openai monorepo, via connected GitHub and Codex integrations.
HOW THIS AFFECTS YOU
●
builderYou must audit all third-party service integrations and SSO flows to prevent lateral movement into core repositories.
●
policyThis highlights the systemic risk of supply chain vulnerabilities in the developer toolchain.