agentleFS
HK-hubHK-hub/AgentSkills/figma-use-motion/SKILL.md

figma-use-motion

Skill6 starsChanged 9 months agoMITView on GitHub

Motion / animation context for the `use_figma` MCP tool — animating Figma nodes via manual keyframes, animation styles, easing, and timeline duration. Load alongside figma-use whenever a task involves adding, editing, or inspecting animation on a node.

4 other projects use a version of this, 2 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Bail out fast on that error: Do not retry; tell the user motion isn't enabled for them and stop.

    line 17
  2. Very specific:

    Size to what you must read: Start small — constraint: { type: 'WIDTH', value: 320 }, quality: "low" — but text and small elements blur …

    line 60

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

use_figma — Figma Plugin API Skill for Motion

Motion context for the use_figma MCP tool. figma-use covers the foundational Plugin API rules — load both together.

Always pass skillNames: "figma-use-motion" (comma-separated alongside figma-use) when calling use_figma for motion work. Logging only.

Runtime Gating

Motion APIs are gated behind the metronome user feature flag. When the calling user doesn't have it, every motion property and helper referenced in this skill throws "<name>" is not a supported API.

Bail out fast on that error. Do not retry; tell the user motion isn't enabled for them and stop. Otherwise you'll burn calls and confuse the user with repeated identical failures.

When to use this skill

Load this skill whenever a use_figma task involves:

  • Adding, editing, or removing keyframes on a node (manualKeyframeTracks, applyManualKeyframeTrack, removeManualKeyframeTrack).
  • Animating fill or stroke colors over time.
  • Applying, editing, or removing animation styles (applyAnimationStyle, removeAnimationStyle, animationStyles).
  • Reading or writing timeline duration via node.timelines / node.setTimelineDuration(id, seconds).
  • Choosing easing for any of the above.

Static design work (creating shapes, components, variables, layout) goes through figma-use alone — this skill is only for the time dimension.

Exposed motion API surface

  • node.manualKeyframeTracks — read/write manual keyframes (including fill, stroke, and effect tracks).
  • node.applyManualKeyframeTrack(field, track) / node.removeManualKeyframeTrack(field) — add, replace, or remove one manual keyframe track without rewriting the whole object.
  • node.animationStyles — read/write animation-style metadata applied to a node.
  • node.applyAnimationStyle(styleId, presetData?) / node.removeAnimationStyle(id) — apply a discovered style and remove an applied style instance by its returned/read-back id.
  • node.timelines — read-only timeline list for the containing top-level frame, with durations in seconds.
  • node.setTimelineDuration(id, durationSeconds) — write the containing top-level frame timeline duration.
  • node.animations — read-only resolved keyframe data (currently manual tracks only — see motion-patterns.md).
  • figma.motion.figmaAnimationStyles() — read-only list of Figma's first-party animation styles.

Authoring custom "figma:motion" preset module source code is out of scope. If the user wants a brand-new animation style, say so and stop; don't fabricate one.

Reference docs

Load these as needed based on what the task involves:

DocWhen to loadWhat it covers
motion-patterns.mdAdding/editing motion animationManual keyframes, animated fills/strokes, applying animation styles, timeline duration
motion-easing.mdSetting animation easingKeyframe easing objects, custom cubic/spring, HOLD, applying easing inside an animation style

Verifying the animation

get_screenshot shows only the timeline's resting state, never motion. To check motion, export_video and sample frames — but it renders server-side and is slow and expensive (~10s to minutes), so make each render count.

Plan before rendering — cost scales with pixels × frames, so keep both no larger than the frames need:

  1. Pick the moments first. You need one frame per phase (e.g. per stagger step, or start / mid / settle), not smooth playback — usually 4–6. This count sets your fps.
  2. Size to what you must read. Start small — constraint: { type: 'WIDTH', value: 320 }, quality: "low" — but text and small elements blur there, so raise WIDTH (768+) when you need to judge fine detail. Omitting constraint = full size (1x; server clamps to 10x / 4096px).
  3. Set fps just high enough to land those frames: fps: 5 covers a handful; 10 is an upper bound. Higher just bloats the render.

Mechanics: export_video works only on a top-level frame whose children carry the animation (pass that frame, not the descendant you keyframed). It returns a jobId with status: "processing" — re-invoke with { fileKey, jobId } to poll. Then extract frames locally with ffmpeg -ss <t> -i anim.mp4 -frames:v 1 frame_<t>.png — extraction is free, so once you've paid for the render, mine it for every frame that tells you something rather than re-exporting. Without a frame extractor like ffmpeg, skip the export and reason about the keyframes instead.

Iterate until it's right. The export is a diagnostic, not a sign-off: if the frames are wrong (bad order, off timing, a missing element, a mask blanking the composite), fix the keyframes/styles and re-export. Read all the frames and batch every fix into one pass before re-rendering — every render carries real overhead, so make each one count instead of re-exporting after each small change.

Skip the export entirely for trivial or self-evident changes.

Pre-flight checklist

In addition to the figma-use pre-flight checklist, verify:

  • Easing uses the public { type: 'EASE_OUT', easingFunctionCubicBezier?: …, easingFunctionSpring?: … } shape — not internal scenegraph names like OUT_CUBIC.
  • Ease-in-out uses the exact public enum EASE_IN_AND_OUT (or EASE_IN_AND_OUT_BACK); never emit the invalid alias EASE_IN_OUT.
  • The node being animated is not a top-level frame (direct child of a page). Animate descendants instead.
  • Timeline values are seconds in the public Plugin API. Extend via setTimelineDuration; never shorten unless the user asked.
  • Transform keyframe fields use public names (TRANSLATION_X, TRANSLATION_Y, ROTATION, SCALE_X, SCALE_Y, SCALE_XY), not internal MOTION_* scenegraph names.
  • Manual keyframe fields come from the public allowlist in motion-patterns.md; generated/internal scenegraph fields intentionally throw.
  • Mutated node IDs are returned (per figma-use Rule 15).
  • When motion correctness isn't self-evident and a frame extractor (ffmpeg) is available, verify via export_video + frame sampling — render small, low fps, iterate until right (see the Verifying the animation section above). get_screenshot shows only the resting state.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.