Skip to main content
skill CX Codex

video-prompting-guide — Claude Code skill: writing AI video prompts

video-prompting-guide breaks a video prompt into shot, movement, light and style, with phrase tables and per-model notes. For the point where plain descriptions stop working.

no license 268 installs ★ 31 updated 2026-08-31

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add 101-skills/superpowers --skill video-prompting-guide -a claude-code
Open on GitHub

What video-prompting-guide does

The skill splits a prompt into slots: shot type, subject, action, setting, lighting, style, technical tags. Each slot has a table of phrases you can lift — low angle, dolly in, rim lighting, film grain. Separate sections handle negative prompts and the usual failures: one vague scene, several scenes crammed together, nothing moving.

Every example calls a model through belt, the inference.sh CLI, as belt app run <model> --input '{...}'. The skill itself renders nothing. Without the CLI installed and a belt login session, those commands just error out. Tool access in the frontmatter stops at Bash(belt *).

The loop it teaches is iterative. Start with a bare scene, tighten the specifics, then add style and technical detail. Each round shows what changed.

What it helps with

  • turn a vague scene idea into a structured prompt
  • pick a camera move that fits the mood
  • fix a clip that comes out static
  • keep lighting and style keywords from clashing
  • move one prompt across different video models

Example prompts for the agent

  • Break down my Veo prompt and tell me what's missing on lighting and camera movement.
  • Write a prompt for a product clip: phone on a white pedestal, soft studio light.
  • My Kling clip looks frozen — show me how to describe motion in the prompt.

How to install video-prompting-guide

1Skills CLI (Vercel)

npx skills add 101-skills/superpowers --skill video-prompting-guide -a claude-code

2Manually: put the files in place

~/.claude/skills/video-prompting-guide/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/video-prompting-guide/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about video-prompting-guide

How to install video-prompting-guide in Claude Code?

Run npx skills add 101-skills/superpowers --skill video-prompting-guide -a claude-code. The skill lands at ~/.claude/skills/video-prompting-guide/SKILL.md.

What does video-prompting-guide do?

It teaches prompt writing for video generators: frame structure, tables of shot types and camera moves, lighting and style keywords, plus notes for individual models.

What do I need before the examples run?

The belt CLI from inference.sh, installed with npx skills add belt-sh/cli, and a session from belt login.

Similar skills

All skills for AI and ML →

From the same set81

The whole set →
p-video p-video generates videos from text prompts or still images using Pruna models via the belt CLI. Handy for creating 720p and 1080p clips directly in your agent terminal. CX Codex +1 14,500installs p-image A skill for Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, and OpenCode that connects four Pruna P-Image models through the belt CLI. Covers text-to-image, image editing, LoRA style presets, and multi-image compositing. CX Codex +1 14,500installs p-video-avatar p-video-avatar — a skill for generating talking avatar videos from portrait images via CLI. Built-in text-to-speech, 30 voices, 10 languages, up to 1080p resolution — 18x faster and 6x cheaper than competitors. CX Codex +1 14,400installs ai-content-pipeline ai-content-pipeline — a skill that chains image, video, voice and music generation into one sequence on top of the inference.sh belt CLI. CX Codex +1 270installs data-visualization data-visualization picks the right chart for your data, sets a palette and axis labels, then renders the figure through the belt CLI. CX Codex +1 255installs prompt-engineering prompt-engineering — a skill that lays out prompt structure for text models, image generators and video models, with examples you can run through the belt CLI. CX Codex +1 251installs explainer-video-guide The explainer-video-guide skill maps an explainer video from script to assembly, with guidance for pacing, scene design, voiceover, and generated visuals. CX Codex +1 249installs logo-design-guide logo-design-guide guides AI agents through logo concept generation and prompt styling via the belt CLI, covering rules from favicon sizes to vectors. CX Codex +1 249installs

Project sponsors

Companies that keep this analytics open