PhiKitA: Phishing Kit Attacks Dataset for Phishing Websites Identification

Recent studies have shown that phishers are using phishing kits to deploy attacks faster, easier and more massive. Detecting in deployed websites might help detect campaigns earlier. To the best of our knowledge, there no datasets providing a set used were attacked by phishing. In this work, we propose PhiKitA, novel dataset contains also generated these kits. We applied MD5 hashes, fingerprints, graph representation DOM algorithms obtain baseline results PhiKitA three experiments: familiarity analysis kit samples, website detection identifying source website. analysis, find evidence different types small campaign. binary classification problem for detection, algorithm achieved an accuracy 92.50%, showing data contain useful information classify Finally, hash obtained 39.54% F1 score, which means does not extract enough distinguish their sources properly.