Couldn't you let Wan make a 5 sec video, then let it continue its train of thought with the prompt, make another 5 sec video, and then stitch the clips together?