ai-video-prompting - SKILL.md Agent Skill

name: ai-video-prompting description: AI video prompt design guide for crafting prompts for Sora, Veo, Kling, Runway, Pika, Luma, Seedance, Wan, image-to-video, reference-based generation, cinematic shots, camera movement, motion continuity, timing, and negative prompts. Use when improving prompts for AI video quality rather than generating still images. source: HsinPu/Autoverse-Ai-Agent-Skills license: Apache-2.0

Use this skill when the task is to write or improve prompts for AI video generation.

Identify the subject, action, setting, duration, aspect ratio, and intended platform.
Specify temporal behavior: starting state, motion path, action beats, ending state, and pacing.
Add camera language: shot size, angle, lens feel, camera movement, framing, and focus behavior.
Add visual direction: lighting, color, texture, era, environment, mood, and production style.
Add continuity constraints for characters, wardrobe, products, logos, props, and scene geography.
Add negative constraints only when they reduce common artifacts or model-specific failure modes.

Subject and action over time, setting, camera/framing, motion and pacing, lighting/style, continuity constraints, output constraints

Describe motion explicitly; still-image prompt language is not enough for video.
Keep one clear action per short clip unless the model supports multi-shot prompting well.
Avoid contradictory camera moves, impossible physics, or too many simultaneous actions.
Prefer concrete shot vocabulary over vague cinematic adjectives.
Preserve the user's creative intent; improve clarity rather than changing the concept.

For still-image prompts, use ai-image-prompt-design.
For selecting models or running a generation workflow, use ai-video-generation.
For multi-shot planning, use storyboard-creation.