image-to-video · 101-skills — Claude Code skill: animate stills into video
A reference skill for turning still images into short videos: model choice, motion prompts and camera moves.
npx skills add 101-skills/superpowers --skill image-to-video -a claude-codeWhat image-to-video does
The skill walks an agent through picking a model and writing a prompt for image animation. It ships a table of models with their app IDs: Wan 2.5 i2v for realistic motion, Pruna's WAN-I2V for fast economical runs, Seedance 2.0 and its fast variant, Fabric 1.0 for cloth and liquid, Grok Imagine Video for general animation. Separate tables cover camera moves (dolly, pan, tilt, orbit, crane) and in-frame motion — water, hair, fog, mechanical parts.
The rule the skill keeps coming back to: subtle motion beats dramatic action. Asking for too much movement produces distortion and artifacts, so prompts should stay calm. There are duration guidelines too — 2-3 seconds gives the best quality, and past 10 seconds quality drops, so longer pieces are meant to be stitched from short clips with infsh/media-merger.
Running it needs the belt CLI from inference.sh and a belt login session. The skill calls it through Bash(belt *), so nothing fires without the CLI installed. Its install steps live in a separate doc, and the README suggests adding the belt-sh/cli skill first with npx skills add belt-sh/cli.
What it helps with
- pick the right model for a landscape, portrait or fabric shot
- write a motion prompt that does not wreck the frame
- add a camera move to a static photo
- stitch a longer video from several short clips
- build a product animation with an orbit around the object
Example prompts for the agent
- animate this photo: slow dolly forward, gentle ripples on the water
- which model should I use to animate fabric on a product shot?
- generate three clips from scene.png with progressive motion and merge them
How to install image-to-video
1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill image-to-video -a claude-code2Manually: put the files in place
~/.claude/skills/image-to-video/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/image-to-video/
1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill image-to-video -a codex2Manually: put the files in place
.agents/skills/image-to-video/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill image-to-video -a cursor2Manually: put the files in place
.cursor/skills/image-to-video/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install 101-skills/superpowers image-to-video2Manually: put the files in place
.github/skills/image-to-video/SKILL.md1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill image-to-video -a gemini-cli2Platform CLI
gemini skills install https://github.com/101-skills/superpowers --path tools/video/image-to-videoManually: put the files in place
.devin/skills/image-to-video/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill image-to-video -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about image-to-video
How do I install image-to-video in Claude?
Run npx skills add 101-skills/superpowers --skill image-to-video -a claude-code. The skill lands at ~/.claude/skills/image-to-video/SKILL.md.
What does image-to-video do?
It tells the agent which model to use for a given image and how to describe camera and subject motion in the prompt. It also gives duration guidance for clips.
What do I need for the skill to work?
The belt CLI from inference.sh and a belt login session. Without it, belt app run commands will not execute.
Similar skills
Everything for AI and ML →image-to-video by other authors comes first.
