Security

Reviews AI-generated code for injection risks, credential exposure, dependency vulnerabilities, and access control gaps—catches what agents miss when they optimize for functionality over safety.