Generative AI video tools have achieved remarkable visual quality, yet directing exact character performance remains a major hurdle. Standard text prompts often struggle to execute precise