A LoRA trained on images of people's shadow with clothes over it. The only needed trigger is dressed shadow.