video-prompting

Generates structured text-to-video prompts for short-form social media marketing videos.

Updated Apr 13, 2026
One-click install
npx skills add https://github.com/rymqan/staple-tech --skill video-prompting-rymqan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: video-prompting
Source: https://github.com/rymqan/staple-tech/tree/main/.claude/skills/video-prompting
Command: npx skills add https://github.com/rymqan/staple-tech --skill video-prompting-rymqan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Writing prompts for text-to-video models like Seedance 2.0 or Runway often produces jittery, inconsistent, or off-brand results because the models require precise camera, lighting, and timing vocabulary. This Skill provides a production-grade prompt engineering framework so marketing teams get usable 5-15 second video clips for social media on the first attempts. ## Core Features & Use Cases - 6-Part Prompt Formula: Structures every prompt as Subject, Action, Environment, Camera, Style, and Constraints, targeting 60-100 words per shot segment. - Hook & Timeline Frameworks: Provides 10 scroll-stopping hook patterns for the first 2 seconds and timeline segmentation rules (e.g., 3-5 segments for a 15-second video) with transition types. - Camera, Lighting & Platform Specs: Supplies a controlled camera movement vocabulary, 8 marketing lighting setups, negative constraint defaults, and aspect ratio/duration rules for Instagram Reels, TikTok, YouTube Shorts, Stories, and LinkedIn. - Use Case: A marketer needs a 15-second Instagram Reel showcasing a new candle product. The Skill produces a complete segmented prompt with a product-materialization hook, warm candlelight setup, dolly-in detail shots, and negative constraints, ready to paste into a video generation API. ## Quick Start Write a 15-second 9:16 video prompt for Instagram Reels showcasing our product with a strong 2-second hook and warm lighting.

Frequently Asked Questions about video-prompting

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I write a prompt for text-to-video models like Seedance or Runway?

Structure the prompt in six parts: Subject, Action, Environment, Camera, Style, and Constraints. Use one present-tense action verb and one camera movement per shot, include a lighting description, and keep each shot segment between 60 and 100 words.

What is the best structure for a 15-second video generation prompt?

Place global subject, style, lighting, and negative constraints first, then split the video into 3-5 timestamped segments such as [0s-3s] and [3s-7s]. Each segment gets one shot size, one camera movement, and one action, with transitions like cut, crossfade, or whip pan between segments.

Why does my AI-generated video have jitter and morphing faces?

Artifacts usually come from combining fast camera movement, fast action, and busy scenes in one segment. Default to slow or static camera moves, let only one element be fast, and always include negative constraints like avoiding jitter, bent limbs, and morphing faces.

What aspect ratio should I use for Instagram Reels and TikTok videos?

Use 9:16 vertical format with a maximum duration of 15 seconds for both Instagram Reels and TikTok. Keep the subject in the center 80 percent of the frame to avoid UI overlays, and make the first 2 seconds a strong visual hook.

How many camera movements can I use in one video shot?

Only one camera movement per shot or segment. Two simultaneous camera verbs produce visual chaos in generation models, so express compound motion as sequential beats, such as a slow dolly-in followed by a gentle pan right for the final 2 seconds.