VIVID: A Visual Harness for Agentic Video Editingwith Wenhan Yu, Zhilong Tan, Weikang Li, Lei Sha1 versionW_epy8umxqv1 · current