OpenAI Codex Security Scans 1.2M Commits, Flags 10,561 Critical Bugs

OpenAI has released Codex Security, an AI‑driven agent that automatically reviewed more than 1.2 million recent code commits across a variety of repositories. In its first pass the system identified 10,561 high‑severity flaws, ranging from insecure deserialization to hard‑coded secrets, and generated remediation suggestions that can be applied directly through ChatGPT.

For security teams, the tool offers a scalable way to surface hidden vulnerabilities before they are exploited, cutting remediation time and shrinking the attack surface of in‑house and third‑party code. Codex Security is now bundled with OpenAI’s Pro, Enterprise, Business, and Education plans, giving defenders immediate access to AI‑powered code review and actionable fixes across their development pipelines.

Categories: AI Security & Threats, Vulnerabilities & Exploits

Source: Read original article