Abstract
Everyday almost 20 million of installs [1], minor-party apps are a main reason for the popularity and obsess of facebook. Disastrously, hackers have registered the possibility of using apps for extending malware and spam. The issue is already valid, as we find that at least 13% of apps in our dataset are malicious. So far, the research center has concentrated on detecting malicious posts and campaigns. In this paper, we ask the question: Given a facebook application, can we figure out if it is malicious? Our main contribution is in evolving FRAppE—Facebook’s Rigorous Application Evaluator—feasibly the first tool focused on detecting malicious apps on facebook. To implement FRAppE, we use data gathered by monitoring the posting action of 111K facebook apps [2] seen across 2.2 million users on facebook. Initially, we determine a set of features that help us classify the malicious apps from favorable ones. For instance, we figured it out that malicious apps frequently share names with other apps, and they generally seek fewer permissions than favorable apps. Secondly, leveraging these characterizing features, we present that FRAppE can encounter malicious apps with 99.5% precision, with no false positives and a elevated true positive rate (95.9%). Finally, we scrutinize the ecosystem of malicious facebook apps and find mechanisms that these apps use to generate. Fascinatingly, we identify that many apps collaborate and assist each other; in our dataset, we identify 1584 apps facilitating the viral proliferation of 3723 other apps across their posts. Long term, we notice FRAppE as a step forward building an independent protector for app appraisal and grading, so as to alert Facebook users before installing apps.