Monitoring web applications for vulnerability discovery and removal under attack

Tese de mestrado, Engenharia Informática (Arquitetura, Sistemas e Redes de Computadores) Universidade de Lisboa, Faculdade de Ciências, 2018Web applications are ubiquitous in our everyday lives, as they are deployed in the most diverse contexts and support a variety of services. The correctness of these applications, however, can be compromised by vulnerabilities left in their source code, often incurring in nefarious consequences, such as the theft of private data and the adulteration of information. This dissertation proposes a solution for the automatic detection and removal of vulnerabilities in web applications programmed in the PHP language. By monitoring the user interactions with the web applications with traditional attack discovery tools, it is possible to identify malicious inputs that are eventually provided by attackers. These in- puts are then explored by a directed static analysis approach, allowing for the discovery of potential security issues and the correction of bugs in the program. The solution was implemented and validated with a set of vulnerable web applications. The experimental results demonstrate that the tool is capable of detecting and correcting SQL Injection and XSS vulnerabilities. In total 174 vulnerabilities were found in 5 web applications, where 2 of these were previously unknown by the research community(i.e., they were ”zero-day” vulnerabilities)