agentleFS
southleftsouthleft/figma-console-mcp-skills/skills/figma-export-tokens/SKILL.md

figma-export-tokens

Skill83 starsChanged 6 months agoMITView on GitHub

Export Figma variables to design token files in DTCG, CSS custom properties, Tailwind v4/v3, SCSS, TypeScript, JSON, Style Dictionary, or Tokens Studio. Use when the user wants to pull design tokens OUT of Figma into code — triggers: 'export tokens', 'export Figma variables', 'generate CSS variables from Figma', 'turn my Figma variables into a tokens.json / Tailwind config / SCSS', 'sync design tokens to code'. Works on ANY Figma plan (reads via the Plugin API, not the Enterprise-only Variables REST API). For the reverse direction (code → Figma) use figma-import-tokens.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read every local variable in a Figma file and emit design-token files.

    line 9
  2. Very specific:

    Confirm scope & format: Ask (or infer) the target format (dtcg is the safe default), whether to split by mode/collection, and the output …

    line 22
  3. Very specific:

    Read the variables: Run scripts/read-variables.js via use_figma (skillNames: "figma-export-tokens").

    line 25
  4. Distinctive:

    Convert — deterministically: Run the bundled converter (Node 18+, zero dependencies).

    line 29
  5. Distinctive:

    Large systems: if the read is huge, scope it to specific collection names before saving.

    line 59

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

figma-export-tokens — Figma variables → design tokens

Read every local variable in a Figma file and emit design-token files. The canonical output is DTCG (W3C Design Tokens Community Group JSON); CSS, Tailwind, SCSS, TS, and JSON derive from it.

Why this beats a raw REST export: Figma's Variables REST API is Enterprise-only (403 on Starter/Pro/Org). This skill reads through the Plugin API via use_figma, so it works on every plan and resolves aliases + multi-mode values that get_variable_defs (default mode only) drops.

Skill boundaries

  • use_figma rules — load the official figma-use skill first; it is the full Figma Plugin API reference. Essentials these scripts rely on: plain JS with top-level await + return (no IIFE, no figma.closePlugin(); console.log is not returned), inputs inlined as const at the top of each script, colors in 0–1 range, load fonts before any text op, await figma.getNodeByIdAsync(...), and atomic errors (a failed script applies nothing — read the error, fix, retry).
  • Reverse direction (code → Figma variables) → use the figma-import-tokens skill.

Workflow

  1. Confirm scope & format. Ask (or infer) the target format (dtcg is the safe default), whether to split by mode/collection, and the output path. Get the file key from the active Figma file or a URL the user provides.
  2. Read the variables. Run scripts/read-variables.js via use_figma (skillNames: "figma-export-tokens"). It returns the normalized collections/modes/variables tree with hex colors, resolved alias references, scopes, and code syntax.
  3. Save the read output to a file, e.g. variables.json.
  4. Convert — deterministically. Run the bundled converter (Node 18+, zero dependencies). Do not hand-write the conversion — this script is the source of truth and produces identical output every run:
    node scripts/convert-tokens.mjs variables.json --format dtcg --out tokens/
    # --format: dtcg (default) | css-vars | tailwind-v4 | tailwind-v3 | scss | ts-module
    #           | json-flat | json-nested | style-dictionary-v3 | tokens-studio
    #           (aliases: css, tailwind, ts)
    # --out <dir>  write file(s) — tokens-studio writes several; omit to print to stdout
    # --prefix <p>            prefix CSS/SCSS var names
    # --modes Light,Dark      include only these modes (default: all)
    # --collection <substr>   include only collections whose name contains <substr>
    This matches the Console figma_export_tokens formatters. It handles exactly what freehand conversion gets wrong: per-type units (opacity/line-height unitless, spacing/radius px); multi-mode output (CSS & Tailwind v4 emit :root + .dark/[data-theme]; TS/JSON emit {mode: value}; SCSS suffixes modes; DTCG keeps them in $extensions); aliases → var()/{ref}; font-weight names → numbers; DTCG round-trip metadata; and it warns on slug collisions and non-numeric weights. (style-dictionary-v3 and tailwind-v3 use the primary mode, matching the Console — those formats have no native multi-mode encoding.)
  5. Report. Surface the written path(s) and any warnings the converter printed (collisions / weight issues are real findings about the Figma file, worth flagging to the user).

Notes

  • The converter is deterministic and authoritative. references/token-formats.md documents the formats it emits; it is reference, not a thing to re-implement by hand.
  • Where it runs: read-variables.js runs anywhere via use_figma. convert-tokens.mjs is Node, so it needs a terminal-capable agent (Claude Code, the Code tab in Claude Desktop, Cursor, Codex, Gemini CLI). In plain Desktop/web chat (no shell), run the converter on your own machine against the saved variables.json, or accept a best-effort inline conversion for DTCG only.
  • Large systems: if the read is huge, scope it to specific collection names before saving.
  • Aliases across collections become var(--…) / {ref} — export all collections together so those references resolve.

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.