OpenAI Patches ChatGPT Data Exfiltration Flaw and Codex GitHub Token Vulnerability

This article covers two critical security vulnerabilities discovered in OpenAI's ChatGPT and Codex that have since been patched.

• •A ChatGPT flaw allowed covert data exfiltration via a hidden DNS-based side channel in the Linux runtime used for code execution, bypassing AI guardrails entirely without user awareness.
• •Malicious prompts or backdoored GPTs could exploit this to silently leak user messages, uploaded files, and sensitive content without triggering any warning dialogs.
• •A separate command injection vulnerability in OpenAI Codex allowed attackers to steal GitHub User Access Tokens by injecting commands through the GitHub branch name parameter in task creation HTTP requests.
• •The Codex flaw enabled lateral movement and read/write access to victims' entire codebases, affecting the ChatGPT website, Codex CLI, Codex SDK, and Codex IDE Extension.