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