A local decoder that turns a Figma .fig design export into a structured JSON tree of pages, screens, and design nodes. Figma is a collaborative design tool, and the decoder works with version 100 or newer files.

Use it to list design pages, inspect a node, trace a Figma screen to Compose code, or audit a local .fig file.

It lets an agent inspect a local design file without opening Figma. This helps answer questions about specific nodes and compare designs with code or product requirements.

How tomdwipo/claude-soul configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to tomdwipo/claude-soul. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/tomdwipo/claude-soul/main/.claude/skills/fig-decode/SKILL.md

Clone the repo

git clone --depth 1 https://github.com/tomdwipo/claude-soul

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/tomdwipo/claude-soul/fig-decode/github.svg)](https://agentmods.dev/skills/tomdwipo/claude-soul/fig-decode)

Your own site

<a href="https://agentmods.dev/skills/tomdwipo/claude-soul/fig-decode"><img src="https://agentmods.dev/badge/skills/tomdwipo/claude-soul/fig-decode/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/tomdwipo/claude-soul/fig-decode"><img src="https://agentmods.dev/badge/skills/tomdwipo/claude-soul/fig-decode.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 78 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,303 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00078	$0.01303
Opus 5	$0.00039	$0.00651
Sonnet 5	$0.00016	$0.00261
Haiku 4.5	$0.00008	$0.00130

Measured 9d ago against content hash ff592e0dd45d, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

fig-decode scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/skills/fig-decode/SKILL.md · 97 lines

How it starts

The opening of the file, as written. The whole thing — 97 lines — stays where its author put it; the contents beside it link to each section on GitHub.

When to use

Invoke this skill when a user:

Provides a path to a .fig file on disk and asks to "look inside" / "explain" / "list pages" / "find a screen".
Pastes a Figma URL like https://www.figma.com/design/<file-key>/...?node-id=NNNN-NNNN and the corresponding .fig is available locally — translate the URL's node-id to NNNN:NNNN (replace - with :) and look it up.
Asks "which screen in code corresponds to Figma node X?" — combine the decoded structure with a grep over the codebase.

Do not invoke this skill if the file is only accessible via the Figma cloud — use mcp__figma__* tools for those (they need an API token and the file's cloud URL).

Requirements

The decoder runs entirely locally:

node 18+ (uses npm install to fetch kiwi-schema and pako on first run)
zstd CLI (brew install zstd on macOS — Figma .fig v100+ uses zstd for the document payload)
unzip (system-default)

How to invoke

The pipeline lives at .docs/tools/fig-decode/. From repo root:

.docs/tools/fig-decode/decode.sh "/path/to/file.fig"
# → prints the output directory it wrote to, defaults to $TMPDIR/fig-decode-<basename>

Pass --full if the user wants the entire decoded message tree (large — every NodeChange with all 590 fields). Default mode writes only the lightweight nodes-summary.json (one record per NodeChange: {guid, parent, type, name, phase}).

How a node-id maps to the output

Figma URLs encode node ids as NNNN-NNNN. The same id appears in the decoded tree as a GUID {sessionID: NNNN, localID: NNNN}. Use the helper:

.docs/tools/fig-decode/scripts/find_node.py 52502-42168
# → prints the matched node, its parent chain back to the canvas/document, and direct children

Without --out, the helper picks the most recently decoded fig-decode-* directory in $TMPDIR.

Pipeline outputs

$TMPDIR/fig-decode-<basename>/
├── canvas.fig          inner Kiwi binary (unzipped from outer .fig archive)
├── meta.json           outer archive metadata
├── thumbnail.png       embedded preview
├── images/             bundled raster assets the design references
├── schema.bin          DEFLATE-decompressed Kiwi schema chunk
├── schema.txt          human-readable schema (`prettyPrintSchema` output)
├── document.bin        zstd-decompressed Kiwi message stream
├── nodes-summary.json  [{guid, parent, type, name, phase}] for every NodeChange
├── stats.json          message/node counts, type histogram, page names
└── message-full.json   only with --full (large)

Read the full file on GitHub · 97 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

9d ago First seen · 97 lines · 78 tokens per session scan A ff592e0dd45d

Subscribe to this mod's changes

fig-decode is a skill published in the GitHub repository tomdwipo/claude-soul (24 stars, last pushed 2d ago), licensed MIT. It adds 78 tokens to every session and 1,303 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Design and UX

webgl-holographic-foil

A self-contained WebGL2 hero: thin-film interference over a crushed-foil surface whose palette shifts with the viewing angle; move the cursor to tilt the film.

nexu-io/open-design · 41 tokens

html-ppt-hermes-cyber-terminal

OpenDesign + BYOK: choosing and wiring your own model, hands-on — cost, quality, and the routing decision. Built as a decision-grade AI literacy deck for engineers, IT, applied-AI teams.

nexu-io/open-design · 53 tokens

html-ppt-taste-brutalist

16:9 HTML deck in tactical-telemetry / CRT-terminal taste. Deactivated-CRT charcoal slides, white-phosphor monospace, hazard-red accent, scanline overlay, ASCII syntax, density over decoration. Distilled from Leonxlnx/taste-skill brutalist-skill (Tactical Telemetry mode).

nexu-io/open-design · 78 tokens

claude-soul: Skill for Claude Code