Sora 2

Use this skill when the user wants actual video generation with OpenAI sora-2, or when the task requires strong prompting, shot planning, and API parameter selection for that model. This skill should unlock operational video work, not just prompt suggestions.

Philosophy: Direct A Scene, Not A Buzzword Stack

Video generation works best when the request reads like direction, not decoration. The job is to translate a vague idea into a concrete scene with clear subject, motion, camera, timing, and constraints.

Before generating, ask:

• What is the deliverable: concept clip, shot test, motion study, image-to-video continuation, or edit/remix?
• What must stay stable: subject identity, framing, shot type, duration, motion tempo, lighting, or style?
• What is the real priority: cinematic mood, product clarity, character continuity, motion readability, or iteration speed?
• Is the user asking for one clip, or a system of related clips with shared continuity?

Core principles:

1. Scene clarity beats adjective piles: subject, environment, action, and camera matter more than long style-word lists.
2. Timing is part of the prompt: duration, beat structure, and camera movement change the usefulness of the clip.
3. Asynchronous generation is part of the workflow: a video job is not done when created; it is done when the job completes and the file is downloaded.