Automating Detection of "Random-Looking" Algorithmic Domain Names

This report presents a practical methodology and accompanying code for detecting algorithmically generated or "random-looking" domain names in passive DNS data. It covers constructing and cleaning English and multi-language wordlists, filtering popular/whitelisted domains and DNSSEC/IDN/ARPA records, extracting effective second-level labels, searching for embedded words, using Shannon entropy and dash-count heuristics, and testing the approach on a 10M-record sample from Farsight SIE Channel 204 (resulting in a reduced candidate set and example outputs); full scripts and supporting tools are provided in appendices.