An agent skill that grades your AI video like a looksmaxxing forum
motionmaxxing turns generic AI motion graphics into films that look decided — and proves it with measured eases and lint gates instead of vibes.

What it does
motionmaxxing is an agent skill (a SKILL.md folder, runs in Claude Code or any agent that reads one) that produces launch films, promos, and kinetic-type pieces from plain HTML, rendered locally with Chrome and ffmpeg. Give it a website and it captures the brand, picks an idea, and animates with timing derived from frame-by-frame analysis of 53 professional motion moments. Crucially, it then renders the film and inspects it with scripts rather than assuring you it came out fine.
The interesting bit Instead of a list of opinions, the skill is built on measurement: ~4,300 decoded pro moves, 12 measured eases, and a “slop autopsy” of rejected AI films that ranked 25 default failure patterns — several of which turned out to be prescribed by the skill’s own earlier wording. The fix was editing the skill, not scolding the agent.
Key highlights
- Concrete numbers a designer can use: median move is 10 frames; pros use 3,392 transform entries against 584 opacity fades; ~96% of moments use hard cuts, cut on motion.
- Scripted gates (G0–G5) plus a page-chrome lint (
lint.mjs) catch the loudest slop tells: corner labels, “02 / 05” counters, typed-out headlines, vibe-coded UI cards. - Honest rating scale — Slop / Mid / Maxxed — and the skill states which notch it actually landed on.
- Before/after comparisons (Cluely, Spotify, Wispr Flow) claim a glow-up, not a miracle; the author notes their blind A/B was one rater, “a signal, not a benchmark.”
Caveats
- Early release, per the project’s own badge; known limits are tracked separately.
- The lint is a lead, not a verdict — it can’t see text inside images or canvas.
Verdict Worth a look if you generate motion graphics with agents and are tired of everything arriving by fade. If you hand-animate in After Effects, the findings table alone is still a nice cheat sheet — the rest you can skip.
Frequently asked
- What is Tejashmakwana/motionmaxxing?
- motionmaxxing turns generic AI motion graphics into films that look decided — and proves it with measured eases and lint gates instead of vibes.
- Is motionmaxxing open source?
- Yes — Tejashmakwana/motionmaxxing is open source, released under the Apache-2.0 license.
- What language is motionmaxxing written in?
- Tejashmakwana/motionmaxxing is primarily written in JavaScript.
- How popular is motionmaxxing?
- Tejashmakwana/motionmaxxing has 696 stars on GitHub.
- Where can I find motionmaxxing?
- Tejashmakwana/motionmaxxing is on GitHub at https://github.com/Tejashmakwana/motionmaxxing.