Prosocial behavior is a distinguishing characteristic of human nature. Although prosocial behaviors emerge early in development, contextual factors play an important role in how these behaviors are manifested over development. A large body of research focuses on the trajectory of prosocial development across diverse cultures and investigating contexts that foster it. Against this backdrop of developmental research endeavoring to understand and enhance the cooperative side of humanity, is the catastrophic impact of profoundly negative forces on social-emotional development for children forced to flee from violent conflict. Close to half a million Rohingya children, whose families were forced to flee genocide in Myanmar, now live in the largest refugee camp in the world. To examine the resilience of human prosociality in the face of extreme adversity, we documented initial levels of prosociality in Rohingya refugee children living in a mega-camp (Cox's Bazar, Bangladesh) and the extent to which those levels were improved following a multifaceted intervention designed to foster prosociality. The research was a partnership between Rohingya community members with lived experience, humanitarian practitioners, and developmental researchers. A sample of 152 Rohingya children (5-12 years) participated in pre- and postintervention assessments of prosocial behaviors and related cognitive-affective processes. The 10-day collaboration-based intervention was implemented between November 2021 and January 2022 by Rohingya researchers. Birthplace was used as a proxy measure of trauma level. Children born in Myanmar (N = 88) directly experienced relatively higher levels of trauma (genocide, forced migration) than children who were born in the camp after their families fled from Myanmar (N = 64). Children were individually tested pre- and postintervention with a task battery, including a helping (Origami) and two sharing tasks (Dictator Game [DG], Forced Choice sharing) measuring prosocial behavior. Assessments of related cognitive-affective processes included measures of empathic responding and emotion perspective-taking in story tasks (Imagine, Judgment) and executive function (EF) skills (Younger: Hearts & Flowers; Older: Dimensional Change Card Sorting). Small group intervention sessions conducted over 10 days targeted these prosocial behaviors and cognitive-affective processes and were based on collaborative activities, emotion perspective taking and EF skills training with the same partner throughout the intervention phase. We used latent change modeling to examine initial levels (preintervention) and intervention-related changes in these measures from pre- to postintervention. Prosocial responding was found across all measures (preintervention) and improvements (pre- to postintervention change) were apparent across most measures. Age and birthplace variables were significant predictors of initial levels and intervention-related change. Initial levels: Regarding age, older children (9-12 years) showed higher levels than younger children (5-8 years) of sharing in the Forced Choice task but lower levels in the DG. Older children also showed higher levels of empathic responding when asked to report how they would feel and respond to another person's misfortune in the Imagine task. Regarding birthplace, prior to the intervention camp-born children showed higher levels than Myanmar-born children of helping in the Origami task and reported more behavioral responses indicating how they would respond to misfortune in the Imagine task. In contrast, Myanmar-born children had higher levels of sharing in the DG and consistently chose equality over inequality in the Forced Choice sharing task, even when their partner would receive more, indicating a pattern of generosity in these children. Myanmar-born children had lower levels than camp-born children on EF measures. Intervention-related change: Regarding age, older but not younger children were more likely to increase choices for equality over inequality on the Forced Choice sharing task following the intervention. Regarding birthplace and helping, camp-born children increased behaviors that helped their partner make origami shapes themselves ("how-to" helping), whereas Myanmar-born children increased behavior that took over folding for their partner ("do-for" helping). For sharing tasks, Myanmar-born but not camp-born children increased sharing in the DG and showed an increased pattern of generosity in Forced Choice sharing task. In the Imagine story task, children born in Myanmar were more likely than those born in camp to increase empathic responding (i.e., imagining how they would feel). Children born in Myanmar showed less improvement on EF measures than children born in the camp. Taken together, these findings provide evidence that in a context of extreme adversity, Rohingya children exhibited prosociality and benefitted from a multifaceted intervention. Our research adds credence to the view that human prosociality is a fundamental characteristic of humanity that not only survives but can be enhanced in even the most adverse of childhood environments. Our multifaceted intervention, which was implemented within a collaborative social context and targeted prosocial behaviors and related cognitive-affective processes, was designed to be easily implemented within existing psychosocial support programs in refugee contexts. As the numbers of children affected by violent conflict and forced migration rise alarmingly worldwide, there is a critical need to expand research partnerships that aim to improve developmental outcomes for these millions of children.