agentleFS
senlindesignsenlindesign/claude2figma/.claude/skills/figma-style-binding/SKILL.md

figma-style-binding

Skill194 starsChanged 7 months agoMITView on GitHub

Triggers on any visual property change in Figma — creating text, setting colors, adjusting spacing/padding/gap/radius. Enforces that ALL values bind to Figma Styles or Variables, never hardcoded. Includes post-write QA verification.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 1 is distinctive, and 5 are 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.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Every visual value must come from the design system.

    line 9
  2. Very specific:

    Every text node must use textStyleId.

    line 28
  3. Very specific:

    If no local style matches, search libraries via search_design_system.

    line 36
  4. Distinctive:

    Every fill/stroke must bind to a COLOR Variable (preferred, supports theming) or Paint Style.

    line 45
  5. Very specific:

    Never use raw { r, g, b } without a binding.

    line 57
  6. Very specific:

    After every use_figma call that creates or modifies nodes, run this verification on the returned node IDs:

    line 92

≈ “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

Style Binding + QA

Every visual value must come from the design system. Supplements figma-generate-design. Prerequisite: figma-preflight must have run this session.


Binding Hierarchy

For any visual property, follow this order. Stop at the first match.

1. Connected Library  →  search_design_system → import → apply
2. Local Style        →  Style Registry → apply by ID
3. Local Variable     →  Variable Registry → apply by ID
4. Gap found          →  Report to user, wait for decision

Text

Every text node must use textStyleId. Individual font properties (fontSize, fontFamily, etc.) are forbidden.

const style = await figma.getStyleByIdAsync("<id>");
await figma.loadFontAsync(style.fontName);
node.textStyleId = "<id>";

If no local style matches, search libraries via search_design_system. If no match anywhere:

⚠️ Text style gap: no style for "[role]". Available: [top 5]. Use closest, or add missing style?

Color Fills

Every fill/stroke must bind to a COLOR Variable (preferred, supports theming) or Paint Style.

// Variable binding (preferred)
const variable = await figma.variables.getVariableByIdAsync("<id>");
const fill = { type: "SOLID", color: { r: 0, g: 0, b: 0 } };
node.fills = [figma.variables.setBoundVariableForPaint(fill, "color", variable)];

// Paint Style binding
node.fillStyleId = "<id>";

Never use raw { r, g, b } without a binding.


Spacing, Padding, Gap, Radius

Bind to FLOAT Variables. layoutMode must be set BEFORE setBoundVariable.

node.setBoundVariable("paddingTop", spacingVar);
node.setBoundVariable("paddingBottom", spacingVar);
node.setBoundVariable("paddingLeft", spacingVar);
node.setBoundVariable("paddingRight", spacingVar);
node.setBoundVariable("itemSpacing", spacingVar);
node.setBoundVariable("cornerRadius", radiusVar);

Spacing can fall back to raw values temporarily with user confirmation. Color and text cannot.


Forbidden / Required

ForbiddenRequired
node.fontSize = 24node.textStyleId = id
node.fills = [{ type: "SOLID", color: { r: .2, g: .4, b: 1 } }]Variable or Style binding
node.paddingLeft = 16node.setBoundVariable("paddingLeft", var)
node.cornerRadius = 8node.setBoundVariable("cornerRadius", var)
Creating a Button from scratchimportComponentByKeyAsync from library

QA Verification

After every use_figma call that creates or modifies nodes, run this verification on the returned node IDs:

const nodeIdsToAudit = [/* paste returned IDs */];
const results = [];

for (const id of nodeIdsToAudit) {
  const node = await figma.getNodeByIdAsync(id);
  if (!node) { results.push({ id, status: "NOT_FOUND" }); continue; }

  const checks = [];

  if (node.type === "TEXT") {
    checks.push({ prop: "textStyleId", bound: !!node.textStyleId });
  }

  if ("fills" in node && Array.isArray(node.fills) && node.fills.length > 0) {
    const bound = !!node.fillStyleId || (node.boundVariables?.fills?.length > 0);
    checks.push({ prop: "fills", bound });
  }

  if ("layoutMode" in node && node.layoutMode !== "NONE") {
    for (const p of ["paddingLeft","paddingRight","paddingTop","paddingBottom","itemSpacing"]) {
      if (p in node) checks.push({ prop: p, bound: !!(node.boundVariables && p in node.boundVariables) });
    }
  }

  if ("cornerRadius" in node && node.cornerRadius > 0) {
    checks.push({ prop: "cornerRadius", bound: !!(node.boundVariables && "cornerRadius" in node.boundVariables) });
  }

  const failed = checks.filter(c => !c.bound);
  results.push({ id, name: node.name, type: node.type, status: failed.length === 0 ? "PASS" : "FAIL", failed: failed.map(c => c.prop) });
}

return { auditResults: results };

If FAIL: Fix each unbound property using the binding rules above, then re-audit. Do not proceed to the next design step until all pass.

Report format:

✅ All [N] nodes passed.
// or
❌ FAIL "Card" (FRAME) — paddingTop, cornerRadius unbound. Fixing...

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.