Introduction: The classic act-person model of shame-proneness defined shame as originating from negative self-appraisals following wrongful actions, conferring broad vulnerability to psychopathology. However, recent developments postulate that shame may originate from real or imagined social evaluation (sociometer view of shame). If so, shame might leave one vulnerable to psychosocial stressors and may manifest in social anxiety specifically, even after accounting for general negative affect. We investigated how shame-proneness predicted concurrent symptoms and prospective responses to interpersonal stressors (social anxiety, feeling evaluated, and other symptoms) over five weeks in a sample including individuals at clinical and subclinical levels of emotional symptoms. Method: We oversampled for individuals meeting criteria for anxiety and depressive disorders based on clinical interviews (n = 58) and also included those not meeting diagnostic criteria (n = 101) to ensure a broad range of symptoms (total N = 159). Participants completed baseline measures of shame- and guilt-proneness, trait negative affect (NA), and symptoms of social anxiety, depression, and generalized anxiety disorders followed by symptom diaries for 5 weeks following their worst psychosocial stressors (1,923 diaries). Results: As expected, even after controlling for NA and guilt-proneness, shame-proneness uniquely predicted concurrent social anxiety (∆R2 = 8%) and prospectively predicted experiences of social evaluation. Shame-proneness demonstrated weaker links to depression, and no unique links to general anxiety and worry. Discussion: Shame-proneness functioned in a manner predicted by sociometer theory, demonstrating specificity for social evaluative symptoms and concerns. Results have implication both for shame theory and clinical practice with shame-prone individuals.