In adversarial networks, edges correspond to enmity or competition. A new type of vertex called a low-key leader is identified in adversarial networks, distinguished by having a high CON-score and low PageRank. We demonstrate the existence of low-key leaders in adversarial networks by considering data from real-world networks. The three types of adversarial networks we consider are: i) dominance networks in animal populations, ii) trading networks between countries, and iii) Bitcoin trust networks. In each case, low-key leaders are detected. We introduce the direct ranking model that generates power-law directed graphs with low-key leaders. We conclude with several open questions for future work.