LAVA - Large-Scale Automated Vulnerability Addition
Explore the innovative LAVA (Large-scale Automated Vulnerability Addition) technique for generating ground-truth vulnerability corpora in this IEEE Symposium on Security & Privacy conference talk. Delve into the challenges of automating vulnerability discovery and the importance of having reliable datasets for evaluating security tools. Learn how LAVA uses dynamic taint analysis to inject realistic bugs into program source code, accompanied by triggering inputs. Discover the application of LAVA to eight real-world programs, including bash and tshark, and examine the preliminary evaluation results of prominent fuzzing and symbolic execution-based bug finders. Gain insights into the potential of LAVA for creating on-demand vulnerability corpora, enabling rigorous tool evaluation and providing high-quality targets for security tool developers.