// #openai
3 articles
OpenAI Rolls Out ChatGPT Lockdown Mode to Block Prompt-Injection Data Exfiltration
OpenAI has released ChatGPT Lockdown Mode, a security configuration that prevents ChatGPT from loading external URLs, rendering images from arbitrary sources, or executing third-party plugin calls — the primary vectors for prompt-injection attacks that cause ChatGPT to exfiltrate data to attacker-controlled endpoints. Enterprise and education customers can now enforce Lockdown Mode organisation-wide via the admin console.
OpenAI Confirms Developer Devices Breached via TanStack Supply Chain Attack — Code-Signing Certificates Rotated
OpenAI confirmed that two developer devices were compromised as a result of the TanStack npm supply chain attack disclosed on 12 May, with malicious postinstall hooks executing on machines running npm install within the six-minute poisoning window. OpenAI rotated all affected code-signing certificates and npm tokens and is investigating whether any internal packages published using the compromised credentials were delivered downstream.
OpenAI Launches Advanced Account Security Programme with Mandatory Phishing-Resistant MFA
OpenAI has announced an opt-in Advanced Account Security programme for high-risk users — journalists, human rights advocates, executives, and researchers — offering phishing-resistant FIDO2 hardware key and passkey authentication, stricter account recovery controls, and session compromise mitigations. The programme, developed in partnership with Yubico, acknowledges that standard MFA is insufficient against sophisticated phishing and AiTM attacks targeting OpenAI accounts with access to sensitive workflows.