AnaDroid: Malware Analysis of Android with User-supplied Predicates

摘要: Abstract Today's mobile platforms provide only coarse-grained permissions to users with regard how third-party applications use sensitive private data. Unfortunately, it is easy disguise malware within the boundaries of legitimately-granted permissions. For instance, granting access “contacts” and “internet” may be necessary for a text-messaging application function, even though user does not want contacts transmitted over internet. To understand fine-grained permissions, we need statically analyze their behavior. Even then, detection faces three hurdles: (1) analyses prohibitively expensive, (2) automated can find behaviors that they are designed find, (3) maliciousness any given behavior application-dependent subject human judgment. remedy these issues, propose semantic-based program analysis, in loop as an alternative approach detection. In particular, our analysis allows analyst-crafted semantic predicates search filter results. Human-oriented systematically, quickly concisely characterize applications. We describe tool provides analysts library ability dynamically trade speed precision. It also inspect details every suspicious state (abstract) execution order make ruling whether or truly malicious respect intent application. addition, permission profiling reports generated aid identifying common behaviors.