An accumulating amount of research has documented the harmful effects of abusive supervision on either its victims or third parties (peer abusive supervision). The abusive supervision literature, however, neglects to investigate the spillover effects of abusive supervision through third-party employees’ (i.e., peers’) mistreatment actions toward victims. Drawing on social learning theory, we argue that third parties learn mistreatment behaviors from abusive leaders and then themselves impose peer harassment and peer ostracism on victims, thereby negatively affecting victims’ performance. Further, we posit that, if a victim has a proactive personality, this will weaken these indirect, negative effects. We conducted two studies, both with three-wave longitudinal data, to verify the hypotheses. The results of Study 1 evidence the significant indirect effects of abusive supervision on employee creative performance via both peer harassment and peer ostracism. Contrary to our moderation hypothesis, the analysis shows that victims’ proactive personality strengthens rather than weakens the negative indirect effects of peer harassment. Study 2 generally replicated the results of Study 1 with employee’s objective job performance as outcome. Our research contributes to the abusive supervision literature by highlighting a social learning process of third-party peer mistreatment, suggesting a spillover channel of abusive supervision on the victim’s performance.