DomXssMicro: A Micro Benchmark for Evaluating DOM-Based Cross-Site Scripting Detection

• Main Authors: Jinkun Pan, Xiaoguang Mao
• Format: Conference Proceeding
• Language: English
• Subjects: Benchmark testing; Browsers; Context; DOM-based XSS; HTML; micro benchmark; Servers; tool evaluation; Uniform resource locators
• ISSN: 2324-9013
• DOI: 10.1109/TrustCom.2016.0065

With the prevalence of JavaScript, Cross-site Scripting based on Document Object Model (DOM-based XSS) has become one of critical threats to client-side Web applications. To detect DOM-based XSS vulnerabilities, a variety of tools have been developed, providing different features and abilities. Both for developers and tool users, the benchmark plays an important role in evaluating the effectiveness of detection tools. However, no widely used standard benchmark exists in the domain of DOM-based XSS. In this paper, we present a micro benchmark named DomXssMicro. DomXssMicro is constructed based on a template extracted from representative vulnerabilities, consisting of six orthogonal components (i.e. Source, Propagation, Transformation, Sink, Trigger and Context). In DomXssMicro, there are 175 test cases in total, each one of which aims at testing a specific property of DOM-based XSS. To illustrate our micro benchmark, an empirical study is performed to evaluate six state-of-the-art DOM-based XSS detection tools, including both commercial and open-source ones. The results demonstrate that our micro benchmark is helpful in providing guidance and insight for tools selection and further improvement.