Detecting and preventing ActiveX API-misuse vulnerabilities in internet explorer

Abstract

ActiveX is used to build reusable software components in Microsoft Windows. It is widely used by many Windows applications, such as Internet Explorer and Microsoft Office. As general-purpose components, ActiveX controls expose methods to applications, which may be used in ways unexpected by the ActiveX designer, leading to malicious activities. We call such misuse of ActiveX methods - ActiveX API misuse vulnerabilities. In this paper, we present a solution which identifies and prevents API misuse of ActiveX controls in Internet Explorer. We construct models to represent normal functionality of ActiveX methods, and identify ActiveX API misuse by identifying the methods that can reach dangerous (system) APIs. We then develop a technique for Internet Explorer to prevent the use of dangerous ActiveX methods. We evaluated our approach on six real-world ActiveX controls. We are able to identify and prevent ActiveX API misuse in these controls. Our approach is effective in detecting ActiveX API misuse and has negligible overhead for preventing attacks. © 2012 Springer-Verlag.