Systems and methods for determining that uniform resource locators are malicious

摘要

A computer-implemented method for determining that uniform resource locators are malicious may include identifying a uniform resource locator that may be posted on a social networking platform and that may be subject to a security assessment, gathering contextual data from the social networking platform that describes at least one instance of the uniform resource locator within the social networking platform, generating, based on the contextual data, a social fingerprint of the uniform resource locator and classifying the uniform resource locator as malicious based at least in part on the social fingerprint. Various other methods, systems, and computer-readable media are also disclosed.

主权项

1. A computer-implemented method for determining that uniform resource locators are malicious, the method being performed by a computing device comprising at least one processor, the method comprising:
identifying a plurality of uniform resource locators that are posted on a social networking platform; gathering contextual data from the social networking platform that describes a plurality of instances of the plurality of uniform resource locators within the social networking platform, the contextual data comprising a number of multiple resource locator masking services used by an instance of a uniform resource locator within the plurality of instances of the plurality of uniform resource locators; generating, based on the contextual data, a plurality of social fingerprints of the plurality of uniform resource locators; determining whether some or all of the uniform resource locators within the plurality of uniform resource locators associated with the plurality of social fingerprints are malicious by determining that some or all of the uniform resource locators within the plurality of uniform resource locators that use above a certain number of multiple resource locator masking services are malicious; generating, based at least in part on the plurality of social fingerprints and the determination, a ground truth set comprising training data for a machine learning system; generating, based at least in part on the ground truth set, a social fingerprint classifier that classifies uniform resource locators that use above the certain number of multiple resource locator masking services as malicious; identifying at least one additional uniform resource locator that is posted on the social networking platform and that is subject to a security assessment; classifying the additional uniform resource locator as malicious using the social fingerprint classifier.