Skip to main content
Agent skill

Video Motion Design

Make a motion design video with Remotion (React) on this server - an intro, a trailer, a product or feature video, a short / reel / story, animated text, a slideshow of screenshots, an explainer. Use when the user asks for a video, an animation, motion design, a short, a reel or a trailer.

All install options
Quick install · Personal (~/.claude/skills)Project / Cursor / Desktop →
mkdir -p ~/.claude/skills/video-motion-design && curl -fsSL 'https://raw.githubusercontent.com/makushingu/harness-maxxing/main/skills/video-motion-design/SKILL.md' -o ~/.claude/skills/video-motion-design/SKILL.md

Get the complete skill folder, including its scripts and reference files.

When to use this skill

Use Video Motion Design when an AI agent needs a reusable SKILL.md workflow for this job: Make a motion design video with Remotion (React) on this server - an intro, a trailer, a product or feature video, a short / reel / story, animated text, a slideshow of screenshots, an explainer. Use when the user asks for a video, an animation, motion design, a short, a reel or a trailer.

When not to use it

Skip Video Motion Design when the task is outside the creativity category, or when a more specific skill in this directory already covers the same workflow with clearer triggers.

How to install

Where will you use it?

Install scope
One-line install (Claude Code)
mkdir -p ~/.claude/skills/video-motion-design && curl -fsSL 'https://raw.githubusercontent.com/makushingu/harness-maxxing/main/skills/video-motion-design/SKILL.md' -o ~/.claude/skills/video-motion-design/SKILL.md
  1. Claude Code personal install: place the skill and its bundled files at ~/.claude/skills/video-motion-design/SKILL.md. For a shared project, commit the folder at .claude/skills/video-motion-design/ instead.
  2. Current Claude Code detects SKILL.md changes in watched directories during the session. If you just created a top-level skills directory, run /reload-skills. Check that the skill is listed before testing its trigger.
  3. For Claude chat or Cowork in the Desktop app, manage uploaded skills in Customize > Skills; copying files into a local Claude Code folder is not a Desktop install. Other agents may use different install locations.

Full install guide for Claude, Cursor, and Codex →

Try a starter prompt

After installing the skill, paste this starter prompt into your agent to verify it triggers on a small task.

Starter prompt
Use the Video Motion Design skill for this task: Use Video Motion Design when this documented workflow matches the task. Confirm any required inputs or files before making changes.

What supports this recommendation?

71/100 provisional content score. This measures documentation, practical detail and source signals. It is not a measured task success rate or a security certification.

Documentation
77/100
Practical detail
68/100
Source evidence
63/100
Source signals
72/100
Model compatibility evidence
  • claudeuntested

    No model-specific signal or recorded compatibility test was found.

  • chatgptuntested

    No model-specific signal or recorded compatibility test was found.

  • geminiuntested

    No model-specific signal or recorded compatibility test was found.

  • copilotuntested

    No model-specific signal or recorded compatibility test was found.

  • llamauntested

    No model-specific signal or recorded compatibility test was found.

  • perplexityuntested

    No model-specific signal or recorded compatibility test was found.

  • mistraluntested

    No model-specific signal or recorded compatibility test was found.

  • grokuntested

    No model-specific signal or recorded compatibility test was found.

Inferred compatibility is based on content signals. Check the app, its available tools and setup requirements before running a skill.

What this skill does

# Motion design videos (Remotion)

Remotion is installed once on the server (`maxxing video status`). A video is a React project: each frame is rendered from components, so everything is code - text, timing, transitions.

## 1. Create the project ```bash maxxing video new <topic folder>/videos/<slug> # e.g. projects/acme/topics/blog/videos/launch ``` It holds `src/Root.jsx` (one `<Composition>` per video: id, size, fps, duration), `src/Intro.jsx` (an example scene), `public/` (images, audio, fonts: use them with `staticFile('logo.png')`). `node_modules` is a link to the shared install: to add an npm package, first replace the link with a real install (`rm node_modules && npm install --no-audit --no-fund <pkg>`), never install into the shared one. If `maxxing video new` says Remotion is not prepared: `maxxing deps update` (as root).

## 2. Write it - Agree on the format first: 1920x1080 (landscape), 1080x1920 (shorts, reels, stories), 1080x1080 (square post); 30 fps; duration in frames (30 = 1 s). - Build scenes with `<Sequence from={..} durationInFrames={..}>`; animate with `interpolate(frame, [a, b], [from, to], {extrapolateLeft: 'clamp', extrapolateRight: 'clamp'})` and `spring({frame, fps})`; fonts with `@remotion/google-fonts/<Font>`; images `<Img src={staticFile(..)}/>`; music or voice `<Audio src={staticFile('voice.mp3')}/>`. - Readability first: big text (>= 60 px in 1080p), max ~8 words on screen, each text visible long enough to read (~0.3 s per word, 1.5 s minimum), safe margins (~5 %), strong contrast. - Screenshots: show them whole and highlight with a frame or an arrow rather than zooming into unreadable crops.

## 3. Check, then render ```bash cd <project> && npx remotion still src/index.jsx <Id> out/check.png --frame=<n> # look at key frames maxxing video render <project> --id <Id> --out out/<name>.mp4 # = npx remotion render, through `maxxing heavy` ``` `maxxing video render` (and `maxxing heavy <command>` for an ffmpeg pass) runs with the background priority: when the processor is full the interface and the sessions go first, the render slows down a little and never stops. Look at 2-3 stills (Read the PNG) before the full render and fix what is off. Rendering takes roughly real time to a few times real time on a small server. Give the user the path of the MP4 (and a still as preview). Keep the project: changing a text later = edit + render again.

## 4. Assembling with ffmpeg (overlays, long edits) A small server has a few GB of RAM and often no swap; a turn is capped at `unit_memory_max` (75 % of RAM by default), and a command past it is killed (`Killed`, exit 137). ffmpeg decodes every `-i` input ahead of time and keeps the frames it cannot use yet in memory, so: - never give it dozens of inputs that start later in the video (one `-i` per PNG sequence / Remotion window, shifted with `setpts=...+T/TB`): 1080p RGBA frames are 8 MB each and fill the RAM in seconds. Put them in **one** input instead: an `ffconcat` list (one `file '.../element-%04d.png'` + `option framerate 30` + `duration` entry per sequence, a transparent PNG for the gaps - with `option framerate 30` on that one too: the list takes the clock of its first file, and a 25 fps first entry stretches every sequence x1.2), read by a `movie=filename=list.txt:f=concat:format_opts=safe\=0` source inside `-filter_complex` (a source filter is pulled frame by frame), then a single `overlay`. Same for a clip placed at the end: `movie=` / `amovie=` rather than `-i`. - keep the video and the audio on the same clock: a video shifted with `setpts=PTS+T/TB` and an audio left at 0 make ffmpeg queue decoded frames until the OOM (shift both, or shift the video back at the end). - before the full render, check the memory on a 30 s range (`systemd-run --scope -p MemoryMax=2G ffmpeg ...` or `/usr/bin/time -v`) and the timing of the whole list (`ffprobe -f concat -safe 0 -i list.txt -show_entries packet=pts_time`: steps of 1/30 s, never going backwards).

Intended uses

  • Use Video Motion Design when this documented workflow matches the task.

Implementation notes

  • Review the source instructions and adapt inputs before running the workflow.

Try it, then tell us how it went

A copied prompt does not tell us whether the task worked. Report your result after checking the output. We collect your tool and outcome, never your files or prompt.