Ratnadira Widyasari

The paper demonstrates that security patch detection models trained solely on publicly reported vulnerabilities (NVD) perform poorly when tested on real-world, unreported 'in-the-wild' patches, suggesting the need for diverse training data.

TitanCA presents a novel, multi-agent LLM orchestration framework that significantly improves vulnerability discovery by reducing false positives and identifying numerous zero-day vulnerabilities.

This paper introduces PoCEvolve, a framework that generates proof-of-concept exploits directly from vulnerability-fixing commits.