OpenAI open-sources Codex Security CLI — defenders race to keep up with AI-powered attackers
openai released codex security cli under apache 2.0 to scan and fix code vulnerabilities via terminal. this automates defense against ai-powered attacks, signaling a shift toward offensive and defensive ai security parity.
openai released the codex security cli, a beta command-line tool available under an apache 2.0 license that automates vulnerability detection and remediation directly from the terminal. based on the internal 'aardvark' project which previously fixed over 3,000 critical vulnerabilities for enterprise customers, the tool requires node.js 22 and python 3.10 and supports bulk scans and ci/cd integration. this release directly challenges anthropic's claude security and highlights the escalating arms race where ai assistance for attackers necessitates ai-powered defenses for developers. the open-source approach allows security teams to customize workflows, though the beta status suggests the feature set is still stabilizing.