Title:Data-Driven Vulnerability Detection and Repair in Java Code

Abstract:Java platform provides various APIs to facilitate secure coding. However, correctly using security APIs is usually challenging for developers who lack cybersecurity training. Prior work shows that many developers misuse security APIs; such misuses can introduce vulnerabilities into software, void security protections, and present security exploits to hackers. To eliminate such API-related vulnerabilities, this paper presents SEADER -- our new approach that detects and repairs security API misuses. Given an exemplar, insecure code snippet, and its secure counterpart, SEADER compares the snippets and conducts data dependence analysis to infer the security API misuse templates and corresponding fixing operations. Based on the inferred information, given a program, SEADER performs inter-procedural static analysis to search for any security API misuse and to propose customized fixing suggestions for those vulnerabilities.
To evaluate SEADER, we applied it to 25 <insecure, secure> code pairs, and SEADER successfully inferred 18 unique API misuse templates and related fixes. With these vulnerability repair patterns, we further applied SEADER to 10 open-source projects that contain in total 32 known vulnerabilities. Our experiment shows that SEADER detected vulnerabilities with 100% precision, 84% recall, and 91% accuracy. Additionally, we applied SEADER to 100 Apache open-source projects and detected 988 vulnerabilities; SEADER always customized repair suggestions correctly. Based on SEADER's outputs, we filed 60 pull requests. Up till now, developers of 18 projects have offered positive feedbacks on SEADER's suggestions. Our results indicate that SEADER can effectively help developers detect and fix security API misuses. Whereas prior work either detects API misuses or suggests simple fixes, SEADER is the first tool to do both for nontrivial vulnerability repairs.