Code Execution Vulnerability in OpenAI Codex CLI by OpenAI
CVE-2025-61260

9.8CRITICAL

Key Information:

Vendor

OpenAI

Status
OpenAI Codex CLI
Vendor
CVE Published:
14 April 2026

What is CVE-2025-61260?

A vulnerability has been found in the OpenAI Codex CLI that allows for arbitrary code execution due to improper handling of Model Context Protocol (MCP) configuration files. Specifically, when users execute commands within a compromised repository, the Codex CLI automatically reads and executes commands from local project files such as .env and .codex/config.toml without seeking user confirmation. This behavior can be exploited by attackers to inject harmful commands, thereby compromising the system.

References

http://openai.com
https://research.checkpoint.com/2025/openai-codex-cli-com...

CVSS V3.1

Score:
9.8
Severity:
CRITICAL
Confidentiality:
High
Integrity:
High
Availability:
High
Attack Vector:
Network
Attack Complexity:
Low
Privileges Required:
None
User Interaction:
None
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.