Skip to content

Figma Use skill (use_figma)

The base skill behind every Figma write: the rules an agent follows when it runs Plugin API scripts in your file through use_figma.

Official · Figma
AuthorFigma
LicenceNo licence file; Figma Developer Terms apply
Sourcefigma/mcp-server-guide · SKILL.md on GitHub
Project websitedevelopers.figma.com/docs/figma-mcp-server
Works inClaude Code · Codex, Gemini CLI, Cursor, Copilot
Read from the source on

The figma-use skill is Figma's official foundation skill for its use_figma MCP tool, which lets an AI agent run JavaScript against the Figma Plugin API inside your file: create and edit frames, text, auto layout, components, variables and styles, or inspect a file's structure. Figma marks it a mandatory prerequisite: the agent loads it before every use_figma call.

It is a rulebook rather than a workflow. It lists the mistakes that make canvas scripts fail (unloaded fonts, colours in the 0–255 range, switching pages the wrong way, forgetting to return node IDs) and how to avoid them, and it points to the task skills that sit on top of it: figma-generate-design for screens, figma-generate-library for design-system assets, figma-use-motion for animation. It needs the remote Figma MCP server and a Figma account.

From the skill's SKILL.md: Use the `use_figma` tool to execute JavaScript in Figma files via the Plugin API.

This is an official skill from Figma. Like any skill it gives the agent instructions and may include scripts it runs, so install it from the source linked here, not from a copy elsewhere.

What it does

  1. Sets the ground rules for every use_figma script: plain JavaScript with top-level await, results sent back with return (console.log is invisible), and every created or changed node ID returned so later calls can find them.
  2. Prevents the common failures: loading a font before any text edit, colours as 0–1 values, fills and strokes reassigned as new arrays, auto layout sizing set only after a node is added to its parent, new top-level frames placed away from the origin.
  3. Explains page handling (switch page at most once per script, and split multi-page work into parallel calls) and the differences between Figma Design, FigJam and Slides files.
  4. Tells the agent which companion skill to load: figma-generate-design for any visible design change, figma-generate-library for reusable components and tokens, the 3D transforms skill for perspective and tilt.
  5. Covers design-system work (components, variables, text and effect styles) with dedicated references, and keeps large reads small by returning IDs and counts first.
  6. Defines error recovery: whether a failed script is safe to retry, or whether the agent must read the canvas first to see what changed.

What you get

  • Nothing by itself: it shapes how the agent writes use_figma scripts. The result is whatever the task asked for, created or edited in your Figma file.

When to use it

  • Any agent task that writes to a Figma file through the Figma MCP server: it is loaded automatically before use_figma.
  • Programmatic edits a designer would find tedious: renaming layers in bulk, binding colours to variables, fixing auto layout, auditing a file's structure.
  • You are writing your own Figma skill and want the base rules the official ones build on.

When not to use it

  • Building whole screens or design-system libraries: use figma-generate-design or figma-generate-library, which load figma-use themselves.
  • Turning a Figma design into code: that only reads the file and is figma-design-to-code's job.
  • You use the Figma desktop MCP server only: it cannot write to the canvas, so use_figma is not available there.

What it needs

  • The remote Figma MCP server (https://mcp.figma.com/mcp) connected and signed in; Figma's docs say writing to the canvas works only with the remote server.
  • A Figma account with edit access to the file. Writing outside your drafts needs a Full seat; writing is free during the beta and Figma says it will later be usage-based and paid.
  • For visible design changes, figma-generate-design next to it (the skill makes it mandatory), and figma-generate-library for library assets. The Figma plugin installs all of them.

Files in the skill

File or folderWhat it is for
SKILL.md21 critical rules, page rules, editor modes, efficient APIs (node.query, node.set, createAutoLayout, inline screenshots), the incremental workflow, error recovery and a pre-flight checklist.
references/gotchas.mdEvery known pitfall with wrong and correct code, starting with the font-loading recipe for text edits.
references/common-patterns.mdWorking script scaffolds: shapes, text, auto layout, variables, components, multi-step workflows.
references/component-patterns.mdVariant sets, component properties, instance swap and discovering existing components.
references/variable-patterns.mdCollections, modes, scopes, aliasing and binding variables.
references/working-with-design-systems/Nine short guides to using and creating components, variables, text styles and effect styles.
references/plugin-api-standalone.d.tsThe full Plugin API typings (about 400 KB), searched for exact signatures rather than read whole.

Licence as found in the repository on 2026-11-15: No licence file; Figma Developer Terms apply.

Example prompts

Type one of these after installing; naming the skill makes sure it is used.

  • In this Figma file, rename layers that still have default names like "Frame 12" to descriptive names based on their content, and list what you changed.
  • Bind all hard-coded blue fills on the Pricing page to our color/brand/primary variable.
  • List the pages, components and variable collections in this file so we can plan a clean-up.
  • Convert the cards in the Features frame to auto layout with 16 px gaps, keeping their current look.

How to install figma-use

Install it with the Figma plugin, or with its companions

figma-use is rarely installed alone. Figma's plugin for Claude Code adds the remote MCP server and every Figma skill at once; after installing, run /mcp and sign in to Figma.

If you copy folders by hand, the commands below also copy figma-generate-design and figma-generate-library, because figma-use sends visible design work and library work to them.

claude plugin install figma@claude-plugins-official
# or, by hand:
claude mcp add --transport http figma https://mcp.figma.com/mcp

Cursor, Gemini CLI and VS Code

Figma's guide installs the same plugin in Cursor (/add-plugin figma in agent chat) and as a Gemini CLI extension (then /mcp auth figma). In VS Code, add an HTTP MCP server with the same URL and use Copilot's agent mode.

/add-plugin figma
gemini extensions install https://github.com/figma/mcp-server-guide

Not for claude.ai uploads or the Claude API

The skill only makes sense with the use_figma tool from Figma's remote MCP server. The Claude API skills sandbox has no network access and we have not verified a claude.ai setup.

A skill is a folder with a SKILL.md file. You install it once; the agent reads its short description at the start of each session and loads the full instructions only when a request matches. Pick the place you use:

Claude Code: from the plugin marketplace

  1. Start Claude Code in a terminal with claude.
  2. This plugin is in Anthropic's official marketplace, which Claude Code adds for you the first time you start it.
  3. Install the plugin. Claude Code opens its details first: check what it will add (skills, hooks, MCP servers), then choose whether it is for you, for everyone on this repository, or for you in this repository only.
    /plugin install figma@claude-plugins-official
  4. Type / and look for /figma:figma-use, or simply ask for the task.

From your shell instead of a session:

claude plugin install figma@claude-plugins-official

The plugin installs as a whole, so other skills come with it, including figma-3d-transforms, figma-code-connect, figma-code-workflows, figma-create-new-file, figma-design-to-code, figma-generate-design, figma-generate-diagram, figma-generate-library, figma-generative-plugins, figma-implement-motion, figma-shaders, figma-spec, figma-swiftui, figma-use-figjam, figma-use-motion, figma-use-slides.

Claude Code: copy the folder into your skills folder

Put the skill's folder in ~/.claude/skills/ to use it in every project on this computer, or in .claude/skills/ inside a repository to share it with everyone who works there (commit it).

git clone --depth 1 https://github.com/figma/mcp-server-guide.git
mkdir -p ~/.claude/skills
cp -R mcp-server-guide/skills/figma-use ~/.claude/skills/
cp -R mcp-server-guide/skills/figma-generate-design ~/.claude/skills/
cp -R mcp-server-guide/skills/figma-generate-library ~/.claude/skills/

Claude Code notices new or changed skills in these folders during the session. If the skills folder did not exist when the session started, run /reload-skills once. Then type /figma-use or just describe the task.

On Windows, run the commands in Git Bash or WSL, or copy the folder into the same place with File Explorer.

Codex, Gemini CLI, Cursor and GitHub Copilot

These agents read the same SKILL.md folders. The one folder all four accept inside a project is .agents/skills/, so this copies the skill there:

git clone --depth 1 https://github.com/figma/mcp-server-guide.git
mkdir -p .agents/skills
cp -R mcp-server-guide/skills/figma-use .agents/skills/
cp -R mcp-server-guide/skills/figma-generate-design .agents/skills/
cp -R mcp-server-guide/skills/figma-generate-library .agents/skills/
AgentProject folderYour user folderNotes
OpenAI Codex
.agents/skills/
~/.agents/skills/
Type $ or run /skills to pick a skill; Codex also chooses one when your task matches.
Gemini CLI
.agents/skills/
.gemini/skills/
~/.agents/skills/
~/.gemini/skills/
Or install from Git with gemini skills install <git URL> --path <folder> --consent. Google replaced Gemini CLI with Antigravity CLI for free-tier and Google One users on 18 June 2026; Antigravity uses .agents/skills/ in a project and ~/.gemini/config/skills/ for you.
Cursor
.agents/skills/
.cursor/skills/
~/.agents/skills/
~/.cursor/skills/
Type / in Agent chat to pick a skill. Cursor also reads Claude Code's .claude/skills/ folders.
GitHub Copilot
.github/skills/
.agents/skills/
.claude/skills/
~/.copilot/skills/
~/.agents/skills/
Works in Copilot CLI, the cloud agent, and agent mode in VS Code, Visual Studio and JetBrains (IDE support depends on the version). Or run gh skill install OWNER/REPOSITORY SKILL (public preview, GitHub CLI 2.90.0 or later).
Only install skills you trust. Anthropic's advice is to use skills only from trusted sources, and to read every file of any other skill before using it: a skill can direct the agent to run commands, read files or send data in ways its description does not mention.

Sources for these steps

Checked against each vendor's documentation on 2026-11-14:

How a safe canvas script looks

Every script is small enough to retry, starts by switching to its page once, loads the fonts it touches, makes its changes, and returns the IDs it created or changed. Related work can be batched into one call when a retry would be harmless; work is split at page boundaries or after a real failure.

Checks come from evidence the script returns (IDs, counts, names, sizes), plus normally one screenshot after composing and one after a visual fix, instead of a screenshot after every step.

Large files without huge replies

Broad reads return only IDs, names, types and counts; details are fetched for chosen IDs in a later call, and each inventory (variables, styles, components, pages) is returned separately. The skill tells the agent to measure a reply's size before returning it rather than silently cutting it.

Questions about figma-use

What is use_figma?

use_figma is the Figma MCP server's general-purpose write tool. The agent sends a JavaScript snippet, Figma runs it against the Plugin API in your file and returns whatever the script returns. It works in Figma Design, FigJam, Slides and Specs files, and only on the remote server.

Do I need figma-use if I already have figma-generate-design?

Yes. figma-generate-design and figma-generate-library both require figma-use before any canvas script, because figma-use holds the Plugin API rules they rely on. Figma's plugin installs them together.

Why does Figma call it a mandatory prerequisite?

Canvas scripts fail in ways that are hard to debug: a text edit throws if its font is not loaded, a colour of 255 is out of range, a page switch with the wrong setter throws, and a script that does not return node IDs leaves the next call blind. The skill front-loads those rules so the agent does not learn them by trial and error.

Can it edit FigJam boards and Figma Slides?

use_figma works in all three editors, but each allows different node types. figma-use covers Figma Design and the differences; for FigJam and Slides, Figma provides the figma-use-figjam and figma-use-slides skills.

Is the figma-use skill free?

The files are free to install but not open source: there is no licence file and Figma's Developer Terms apply. You need a Figma account, a Full seat to write outside your drafts, and Figma says writing to the canvas is free during the beta and will later be a usage-based paid feature.

More AI skills

Related skills

Skill categories