Figma Layout Audit for Web
Proves a web page matches Figma: a snippet reads every visible text from computed styles in the page, scripts check them and the images.
$19
Works with the AI Agents you already use
Figma Layout Audit for Web
Example session with this skill installed
Check the /pricing page against the Figma frames 12:340 (desktop 1440) and 12:900 (mobile 390).
- Read your context and instructions
- Compiled the figma layout audit
- Generated the UI component
mismatch pricing@1440 :: body/main/h1#title [Buy now] size 44 != 48; fill #FFFFFF != #FFF4D6
ambiguous pricing@1440 :: body/main/a#continue [Continue] 100:215: 36 #FFFFFF #1A3C7A/4; 100:231: 28 #9AA4B2 -/None
texts 23: ambiguous 1, mismatch 1, ok 21; dynamic without a check 0
problem public/img/pricing/badge.png 200x90 cropped: PNG 200x90 is smaller than the Figma render bounds 212x104
png 6: ok 5, problem 1
Fixed the title size and color, pointed the Continue button at layer 100:215, re-exported the cropped sprite by its render bounds. Second run: texts 23: ok 23; png 6: ok 6. Both reports exit with code 0.
badge.png
TSX · React component
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Design QA for web UI that ends in two reports with exit code 0 instead of "looks fine to me". Framework-independent: the values are computed styles of the rendered page, after the cascade and CSS transforms.
How it works
- Two read-only scripts dump the text layers and layer sizes of the Figma frames through the official Figma MCP server.
collect_web.jsruns in the page (a browser tool of the agent, Playwright, Puppeteer or the console), changes nothing and returns one row per visible text: computed font size, color, text stroke, position,data-i18nkey.audit_labels.pymatches every text to a Figma layer and compares size (3% tolerance), fill, outline color and outline width.audit_pngs.pychecks changed PNG files for cropped art, opaque edges and sizes that no Figma layer has, including@2x,2.0x/and Android density variants.- Every row that is not
okis fixed or gets a recorded decision (who, when, why) inmap.json. Both reports must exit with code 0. compare_element.pycompares composite elements (button frame, plate, text) with the Figma frame pixel by pixel and builds an app / Figma / difference panel.
What you get: SKILL.md with the 9-step order and the traps, reference docs for formats and statuses, the in-page collector with a Playwright example, Python scripts with no dependencies besides Pillow. Texts filled by code are flagged until they are checked with 3 or more value lengths.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse All Claude Skills.
Needs the official Figma MCP server with view access to the file. PNG only: JPG, WebP, vector files and 9-patch are not checked. Gradient fills are compared by size only. Shadow DOM, iframes, SVG text, canvas, pseudo-element content and form field values are not collected. One screen per breakpoint.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 4 days ago
- Accounts
Figma MCP server
- Software
- Python 3.8+ with Pillow
GitA way to run JS in the page
Python
Permissions
File Scopes
The scripts make no network requests. Figma is read through your own Figma MCP connection.