Virtue ethicists often appeal to practical skill as a way of understanding the nature of virtue. An important commitment of a skill account of virtue is that virtue is learned through practice and not through study, memorization, or reflection alone. In what follows, I will argue that virtue ethicists have only given us half the story. In particular, in focusing on outputs, or on the right actions or responses to moral situations, virtue ethicists have overlooked a crucial facet of virtue: namely, that through practice, virtuous agents develop a cache of perceptual skills that allow them to attend to, detect, and identify the relevant features of a perceptual array, the selection of which is central to recognizing and categorizing a situation as a moral situation of a particular type. In order to support this claim, I will appeal to empirical studies of motor expertise, which show that an expert's capacity to attend to and recognize relevant perceptual inputs differs in important respects from the layperson's. Specifically, I will argue that performing the right action in the right circumstances improves an agent's ability to attend to and identify the morally relevant features of a moral situation.