Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Generates production-ready prompts for Seedance 2.0, ByteDance's AI video model. Uses a six-layer architecture (Subject, Action, Camera, Style, Audio, Constraints) with templates for cinematic scenes, multi-shot narratives, and action sequences. Includes IP translation rules for converting branded or proprietary content into moderation-safe visual descriptions.
The six-layer structure prevents prompt drift and the built-in negative checklist reduces common artifacts without repeated trial-and-error.
When to use it
- Writing a cinematic chase sequence prompt with beat-by-beat camera moves
- Generating an action scene with physics-accurate motion and environmental anchoring
- Converting NFT or branded IP names into moderation-safe visual descriptions
- Building a multi-shot narrative with consistent character identity across shots
- Iterating on a video prompt by adjusting one layer at a time before extending duration
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Create a prompt for a wooden creature walking to a cliff edge at sunset
- 1Identifies this as a T2V cinematic scene and selects the cinematic scene template
- 2Defines subject with visual specifics: bark texture, glowing amber eyes, green gems, flowing purple cape, chibi proportions
- 3Structures action as chronological beats: walks to edge, pauses, turns head slowly, cape billowing
- 4Specifies camera: medium shot, slow dolly-in, eye level, 50mm lens feel, single verb per shot
- 5Adds style anchor, golden-hour lighting, ambient audio (wind, no music), and constraints (no extra characters, 10s duration, no morphing)
A paste-ready Seedance 2.0 prompt with 4 negative constraints, 16:9 aspect ratio suggestion, and recommendation to test at 5s before extending
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.