Dependency Risk Auditing for LLM-Generated Projects via Software Supply Chain Analysis

LLM-generated code often pulls external dependencies without evaluating their security posture. We present an automated supply-chain auditing system that inspects dependency versions, known CVEs, integrity guarantees, and transitive dependency trees. Applied to 500 LLM-generated mini-projects, the audit system identifies risky dependencies in 72% of cases and provides secure alternatives. Integrating the system into LLM workflows decreases usage of vulnerable packages by 67%. This demonstrates that pairing LLM code generation with supply-chain intelligence is critical for preventing inadvertent security issues.