agentsclimarketplace
Sign in

Deep component figma

Skill southleft/skills-for-figma/skills/deep-component-figma

Extract a single Figma component (or any node) as a full recursive tree to N levels deep — every child's layout, fills/strokes/effects, typography, resolved bound variables (token names + code syntax), INSTANCE mainComponent references, prototype reactions, and annotations. Use when you need maximum-fidelity data to generate production code for ONE component — triggers: 'get the full component tree', 'deep extract this component', 'give me everything about this node for code gen', 'resolve all the tokens in this component', 'what instances does this component nest', 'extract reactions/prototype links', 'high-fidelity component spec'. Resolves bound variables to token names on ANY plan. Far deeper than the native MCP's get_design_context / get_metadata.From its SKILL.md

Install
npx -y skills add southleft/skills-for-figma --skill deep-component-figma

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 12 stars12 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.8 KB, 996 tokens by cl100k_base, as published. Nobody here has run it

deep-component-figma — unlimited-depth tree with resolved tokens

Walk one node (usually a COMPONENT / COMPONENT_SET / INSTANCE, but any node works) recursively to a depth you choose, capturing at every level: layout (auto-layout, padding, spacing, sizing), visual properties (fills, strokes, stroke weight, corner radius, effects, opacity), typography for text nodes, boundVariables resolved to token names + code syntax, INSTANCE main-component references (including whether the main is a variant and its set), prototype reactions, and dev-mode annotations. This is the richest single-component read for high-fidelity code generation.

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).
  • Whole-file inventory (all tokens/components/styles at once) → use design-system-inventory-figma.
  • Variant set as a CSS state machine (per-state diffs, pseudo-class mapping) → use analyze-component-set-figma.
  • This depth of tree + resolved tokens + reactions is not what the native MCP's get_design_context (rendered single-selection summary) or get_metadata (flat structure) return — it is a design-system code-gen capability layered on top.

Workflow

  1. Get the node id. From selection (figma_get_selection), search, or a pasted id.
  2. Choose a depth. Set const DEPTH (default 10). Deep design components rarely exceed ~6 levels; use a smaller number first to keep the payload manageable, then go deeper if needed.
  3. Run scripts/deep-component.js via use_figma (skillNames: "deep-component-figma"). Set const NODE_ID and const DEPTH at the top.
  4. Generate code. The returned tree mirrors the DOM you should build. Prefer boundVariables token names (and their codeSyntax) over raw hex; use mainComponent refs to reuse imported components; use reactions to wire up interaction/navigation.
  5. If the payload is too big, lower DEPTH (nodes past the limit are summarized with a childCount + _depthLimitReached marker) or target a smaller sub-node id.

What each node carries

  • Identity: id, name, type, visible. Hidden non-component nodes are marked _hidden and not recursed.
  • Layout: layoutMode, axis sizing/align, padding*, itemSpacing, counterAxisSpacing, layoutWrap, min/max width/height, clipsContent.
  • Visual: fills, strokes, strokeWeight, cornerRadius, effects, opacity (only when non-default; figma.mixed values are skipped safely).
  • Typography (TEXT): characters, fontSize, fontFamily/fontStyle, fontWeight, lineHeight, letterSpacing, alignment, truncation, case, decoration.
  • Tokens: boundVariables resolved to { id, name, collection, resolvedType, codeSyntax } per bound property.
  • Instances: mainComponent (id, name, key, isVariant, set name/id) + componentProperties.
  • Definitions (COMPONENT/COMPONENT_SET): componentPropertyDefinitions, variantProperties.
  • Prototype: reactions (trigger + action: navigation/transition/destination).
  • Dev: annotations (label markdown + categorized properties).

Notes

  • Resolved tokens work on ANY plan — the variable lookup is built via the Plugin API, so even non-Enterprise files get token names instead of opaque VariableID: aliases.
  • The root node echoes _variableMapSize and _maxDepthUsed for sanity-checking.
  • Returns affectedNodeIds: [NODE_ID] even though this is a read — keeps the convention uniform.

What ships with it: 1 file

10.1 KB alongside SKILL.md, 1 of them executable

scripts/

Gives 0 of the 12 instructions most plan spec skills give in 996 tokens

Counted across 1,360 of the 2,617 authors here whose files we hold, read 2026-09-06

  • Ask one question at a timein 73 of 1360
  • Write the spec using the templatein 22 of 1360
  • Ask clarifying questions if neededin 19 of 1360, across 18 files
  • Wait for user confirmation before proceedingin 19 of 1360
  • Save plans to the plans directoryin 17 of 1360, across 13 files
  • Check for product marketing context firstin 16 of 1360, across 5 files
  • Read the plan file completelyin 16 of 1360
  • Order tasks by dependencyin 16 of 1360
  • Gather context from the conversationin 15 of 1360, across 9 files
  • Explore the codebase instead of askingin 15 of 1360, across 13 files
  • Wait for explicit user approvalin 14 of 1360, across 13 files
  • Quiz the user on the breakdownin 13 of 1360, across 7 files

Said here and by no other author read

  • Load the official figma-use skill first
  • Run deep-component.js via use_figma
  • Lower depth if the payload is too big
  • Get the node id from selection or search
  • Choose a depth

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.