The document discusses techniques for detecting obfuscated malicious JavaScript campaigns through scalable behavioral analysis. It proposes analyzing runtime behaviors like global variable names, network traffic patterns, and UI interactions to identify malicious scripts, even when code has been packed or obfuscated. Examples are given of how behavioral signatures could detect encrypted cryptocurrency miners and clickjacking scripts by looking for behaviors like cryptomining initialization and dynamic UI hiding/showing. The document argues this behavioral approach can effectively detect variations of known malware families and emerging malicious campaigns.
Related topics: