OpenAI Launches Codex Security that Discover, Validate and Patch Vulnerabilities

OpenAI announced Codex Security, an AI-powered application security agent available in research preview that builds project-specific threat models, validates findings via sandboxed proof-of-concept execution, and generates contextual patches. The announcement highlights that Codex Security scanned large numbers of commits during beta, identified hundreds of critical and thousands of high-severity issues, and helped assign multiple CVEs for widely used open-source components; organizations are advised to review vendor advisories and apply validated patches.