This study investigates whether primary and secondary variants of psychopathy can be identified in an applied, forensic setting based on self-reports of psychopathy and anxiety. Data were available for two samples of detained boys (Sample A: N = 847, Sample B: N = 749). Using three psychopathy dimensions and anxiety as clustering variables, latent profile analysis arrived at 4 latent classes (LCs) that were tentatively labeled as control (LC1), high anxiety (LC2), moderate psychopathy (LC3), and high psychopathy (LC4). Boys in LC4 engaged in higher levels of alcohol/drug use, conduct problems, reactive and proactive aggression than their counterparts in LC1 and in higher levels of conduct problems, alcohol/drug use, and proactive aggression than boys in LC3. Findings further indicated that the risk for future nonviolent arrests was the highest in LC4 as compared with LC2 and LC3, though no class differences in risk for future violent arrests emerged. Overall, these findings were well replicated in Sample B. Exploratory analyses included additional measures of negative affect (depressed feeling and anger–irritability), maltreatment, and/or number of past arrests (as proxy of a 4th psychopathy dimension) as clustering variables and identified all but 1 (LC3) of the 4 aforementioned LCs. Notwithstanding that our findings challenge the expected relevance of differentiating primary and secondary variants of youth psychopathy, they do suggest that it is possible to identify detained boys with high levels of psychopathic traits who display features associated with adult psychopathy. Implications for theory, research, and practice are discussed.