Dynamic phishing content using generative grammars

Phishing prevention and detection algorithms depend on content exemplars to train on in order to effectively identify threats. Developing these exemplars can either be done by hand, which is time consuming and expensive, or taken from attacks that have already been detected in the wild, which limits the ability to detect new or novel threats. In this paper, we describe PhishGen, a system that uses generative grammars to create dynamic e-mail contents for use as test cases for anti-phishing research. In addition, we demonstrate our system's ability to adapt to existing filters in order to ensure the delivery of e-mails without the need to white-list, which provides an additional level of realism for phishing attacks during penetration testing.