How this Agent works
Builds original fashion-film visuals around musical rhythm, styling, and spatial design — canvas, visual movements, performance storyboard, and keyframes.

Skill · v1.1.0
Video creation Skill
Choreographs styling, beats, and camera movement into a high-gloss fashion music film
Builds original fashion-film visuals around musical rhythm, styling, and spatial design — canvas, visual movements, performance storyboard, and keyframes.
You have a track, an artist, or a brand visual direction, and need an editorial-grade yet producible MV pre-production plan.
通过 canvas_read 获取当前画布与 revision,识别已有角色、造型、场景和镜头。
建立序章、推进、爆发、留白和收束的视觉曲线,为每段分配动作密度、镜头速度、色彩和材质变化。
设计原创视觉母题,使服装轮廓、身体动作、置景和摄影共同表达音乐,而非堆砌无关奢华元素。
拆分为可拍镜头:明确景别、焦段意图、机位、表演动作、灯光变化、转场接口和连续性锚点。
展示最小画布变更并取得一次性批准后写入;对选定造型与核心段落逐项申请关键帧生成,随后查询真实任务状态。
**研究校准**:研究歌曲类型的表演语法、灯光变化、造型功能和制作方法,不复制具体 MV treatment 或造型。
The Agent returns its execution process, real Yiluna Tool results, and a final summary; until a generation job finishes, a waiting state is never packaged as a deliverable.
Currently delivers: Planning document, Keyframe images.The final video file is not included yet.
Illustrative still — not an actual run output