OpenAI Codex GUI Vulnerable to Remote Code Execution via Argument Injection
A critical vulnerability in OpenAI Codex GUI allows remote attackers to execute arbitrary code by tricking users into opening malicious files.

A significant remote code execution (RCE) vulnerability has been disclosed in OpenAI Codex GUI, identified by the Zero Day Initiative as ZDI-26-968 and assigned CVE-2026-110930. This flaw, which carries a CVSS score of 7.8, permits attackers to execute arbitrary code on vulnerable systems, though it requires user interaction.
The vulnerability stems from an issue within the handling of the OpenProjectInCodex context menu command. Specifically, the application fails to properly validate user-supplied strings before incorporating them into command-line arguments. This oversight allows a malicious actor to inject commands that are then executed by the system, leading to code execution in the context of the logged-in user.
Exploitation of this vulnerability requires an attacker to trick a user into opening a specially crafted malicious file. Once the user interacts with this file, the argument injection flaw is triggered, allowing the attacker to achieve their objective. The impact can range from data theft to full system compromise, depending on the privileges of the affected user.
OpenAI has acknowledged the vulnerability and has released an update to address the security concern. Users of OpenAI Codex GUI are strongly advised to apply the available patch as soon as possible to mitigate the risk of exploitation. Further details regarding the fix can be found in OpenAI's official security advisory.
The disclosure timeline indicates that the vulnerability was initially reported to the vendor on October 21, 2026, followed by a coordinated public release of the advisory on November 13, 2026. The advisory was updated on the same day, suggesting a rapid response from OpenAI after the initial report.
This vulnerability was discovered and reported by Ikotas Labs, Inc., with their handle @ikotas00. The Zero Day Initiative, a program that buys and discloses zero-day vulnerabilities, facilitated the responsible disclosure process.
While the specific attack vectors and widespread exploitation are not detailed, the nature of argument injection vulnerabilities often means they can be chained with other exploits or used in targeted attacks. The requirement for user interaction means that social engineering tactics could be employed to deliver the malicious file.
This discovery highlights the ongoing security challenges in complex software applications, even those developed by leading AI organizations. Developers must maintain rigorous input validation and secure coding practices to prevent such critical vulnerabilities from impacting their users.