OpenAI Codex Security Scanned 1.2 Million Commits and Found 10,561 High-Severity Issues

**OpenAI has begun rolling out Codex Security**, an AI-powered security agent (in research preview) that scans repositories to identify, validate, and propose fixes for software vulnerabilities. During beta scans it analyzed more than 1.2 million commits, reporting 792 critical and 10,561 high-severity findings across open-source projects (including several CVEs for GnuPG, GnuTLS, GOGS, Thorium and others); the agent builds project context, validates issues in sandboxed environments to reduce false positives, and suggests remediation to minimize regressions.