Objective: Although the Unified Protocol contains multiple distinct skills to target anxiety, depression, and related conditions, researchers have yet to establish if patients' use of these skills contributes to symptom change. Using data from the first-stage randomization of a sequential multiple assignment randomized trial, we tested whether general skillfulness, defined by skill knowledge, frequency, quality, and effectiveness, predicted within- or between-person changes in anxiety and depression. We further tested whether use of particular UP skills (e.g., mindfulness, behavior change) predicted changes in these outcomes. Method: Participants (N = 70; Mage = 33.74, 67% female, 74% white) completed six sessions of core UP modules and reported on their skillfulness and specific skill use using two novel measures, along with their symptoms of anxiety (OASIS) and depression (ODSIS) before each session, and their alliance (WAI) after each session. We disaggregated between- from within-person effects to test session- to-session effects of skillfulness, skill use, and the alliance on symptom change. Results: Greater within-person skillfulness predicted decreases in anxiety and depression. More frequent within- person use of all UP skills predicted decreases in anxiety, whereas more frequent within-person use of alternative actions specifically predicted decreases in depression. Conclusions: This pattern of results provides guidance on the specific aspects of skill training to prioritize in transdiagnostic treatments depending on whether therapists are targeting anxiety or depression.