Prbl is an AI-first code security scanner that detects which files and lines were generated by models, scans specifically for the classes of vulnerabilities AI coding tools commonly introduce (hardcoded secrets, injection, missing auth, fallback secrets, timing-unsafe checks), and uses an automated rewriter that proposes minimal, behavior-preserving fixes validated against a baseline and a rescan; use it alongside your existing SAST to find and safely remediate issues that general scanners miss in AI-generated code, with CI/GitHub integration and PR comments to streamline fixes.