What problem does it solve? Writing effective Seedance 2.0 video prompts requires film-direction judgment most users lack, and vague adjective-stacked prompts produce generic, unstable, or off-brief results. This skill routes every request through a structured operating loop that converts a scene's intent into concrete camera, lighting, motion, and sound choices, while keeping multi-clip stories consistent and platform facts source-dated. ## Core Features & Use Cases - Directed prompt generation: A directing engine reads a scene's dramatic function and derives one coherent setup instead of stacking empty quality words, with 33 worked genre examples. - Multi-clip sequence management: Stateful project tracking, continuation from accepted footage, continuity QC, and scene-level re-anchoring keep long stories coherent across clips. - Safety and multilingual coverage: IP-safe rewrites for protected characters, brands, and real people, plus native prompt vocabularies and examples in Chinese, Japanese, Korean, Spanish, and Russian. - Use Case: A creator wants a five-clip short drama on Dreamina. The skill interviews for intent, plans scenes, writes Clip 01, observes its actual ending, and writes each continuation from the real final frame while preserving reference tags like @Image1 exactly. ## Quick Start Tell your agent: use seedance-20 to write a 5-second Seedance prompt for a woman reading a letter who realizes something, with one slow push-in and near-silent audio.
