Static detection of access control vulnerabilities in web applications

• Authors: Fangqi Sun;Liang Xu;Zhendong Su
• Venue: SEC'11 Proceedings of the 20th USENIX conference on Security
• Year: 2011

Access control vulnerabilities, which cause privilege escalations, are among the most dangerous vulnerabilities in web applications. Unfortunately, due to the difficulty in designing and implementing perfect access checks, web applications often fall victim to access control attacks. In contrast to traditional injection flaws, access control vulnerabilities are application-specific, rendering it challenging to obtain precise specifications for static and runtime enforcement. On one hand, writing specifications manually is tedious and time-consuming, which leads to non-existent, incomplete or erroneous specifications. On the other hand, automatic probabilistic-based specification inference is imprecise and computationally expensive in general. This paper describes the first static analysis that automatically detects access control vulnerabilities in web applications. The core of the analysis is a technique that statically infers and enforces implicit access control assumptions. Our insight is that source code implicitly documents intended accesses of each role and any successful forced browsing to a privileged page is likely a vulnerability. Based on this observation, our static analysis constructs sitemaps for different roles in a web application, compares per-role sitemaps to find privileged pages, and checks whether forced browsing is successful for each privileged page. We implemented our analysis and evaluated our tool on several real-world web applications. The evaluation results show that our tool is scalable and detects both known and new access control vulnerabilities with few false positives.