style guide runtime check

    1

    Audit rendered web app styles against design tokens using a headless browser to catch runtime visual regressions.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    style guide runtime check

    Example session with this skill installed

    Audit the styles of the local dev server at http://localhost:3000 against the tokens in design-tokens.json. Make sure to check the mobile view.

    • Read your context and instructions
    • Compiled the style guide runtime
    • Generated the UI component

    I have completed the runtime style audit for the application. The self-test passed, and the canary was successfully flagged, ensuring the audit results are valid.

    Style check: Local Dev Server

    Coverage: 2 screens x 2 viewports, 48 element checks. Canary flagged. Not covered: hover states, ::before/::after pseudo-elements.

    Findings (3 distinct values, 8 total hits)

    PropertyValueNearest tokenElementsScreensGuide rule
    color#343434#333333 (gray-900).card-text, .footer-linkHome, DashboardText must use gray scale
    fontSize13px14px (text-sm).sidebar-labelDashboardMinimum legible size is 14px
    backgroundColor#f9f9f9#ffffff (white).header-bgHomeUse standard background scale

    Exceptions applied:

    • main > .hero-blob: Sanctioned brand illustration color deviation (DESIGN.md p.4).
    • .legacy-widget: Approved technical debt for deprecated component.

    Assertions: 4/5 passed Console errors: 0

    Next steps

    • Update .card-text in styles/cards.css to use the var(--gray-900) token.
    • Review the 13px sidebar labels; the design guide mandates a 14px minimum for accessibility.
    • I have preserved style-check.config.json for future repeatable runs.

    cards.css

    TSX · React component

    Generated

    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

    Static CSS reviews miss the reality of inherited values, media queries, and dynamic states controlled by JavaScript. This skill audits what the browser actually renders by checking a running web app against its design tokens in a headless browser environment. It catches off-scale font sizes, stray hex codes, and inconsistent border radii that linting and static analysis overlook.

    What it does

    • Runtime style auditing measures computed styles in a real Playwright/Chromium instance.
    • Token validation compares rendered colors, fonts, and spacing against your style guide scales.
    • State-aware checking navigates through modals, tabs, and logged-in views using custom steps.
    • Visual inventory creates a full report of all style values in use when no guide exists.
    • Exception handling applies sanctioned deviations based on design documentation to reduce noise.
    • Cross-viewport testing verifies style consistency across mobile and desktop breakpoints.

    How it works

    1. Self-test runs a diagnostic to verify the local Playwright environment and script integrity.
    2. Config generation builds a JSON map of your design tokens, exceptions, and UI screens.
    3. App execution connects to a local dev server or URL to inspect the live DOM.
    4. Audit run executes the style-check script to flag violations and nearest-token matches.
    5. Report delivery provides a detailed breakdown of findings, element hits, and coverage gaps.

    Frameworks & tools

    This tool requires Node.js 18+ and Playwright with Chromium. It works with any web framework (React, Vue, Svelte) and supports style guides defined in Tailwind, CSS variables, or JSON tokens.

    Why this beats prompting it yourself

    Manual audits are prone to human error and miss styles tucked inside complex component states. This skill uses a specialized runtime script to ensure 100% accuracy in style measurement, catching the exact pixel and hex values that LLMs cannot "see" through code alone.

    Use cases

    • Design system QA to ensure new components strictly adhere to defined tokens.
    • Pre-release regression to catch accidental style drift before merging to main.
    • Legacy cleanup by inventorying every unique style value in a guide-less project.
    • Responsive audit to find font-size violations on specific mobile breakpoints.

    Known limitations

    Does not inspect pseudo-elements or text drawn inside Canvas/SVG. Interactive states like hover or focus are only checked if explicitly triggered via configuration steps.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Frontend & Web Apps.

    Requires Node.js 18+ and Playwright. Does not inspect pseudo-elements or Canvas/SVG text. Hover/focus states require manual configuration.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 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

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 13 days ago

    • Free with an Agensi account

    Software
    PlaywrightNode.js

    Permissions

    Terminal / Shell
    Browser
    Write Files
    Listed13 days ago

    What's inside

    Frequently Asked Questions

    Static linters only read code, whereas this skill uses a headless browser to measure 'computed' styles. It catches issues that only appear at runtime, such as incorrect inherited values, media query overrides, or dynamic styles injected by JavaScript.

    The skill includes the runtime auditing scripts, a configuration template for your design tokens, and the execution logic for Playwright. You also receive a reporting module that categorizes style violations and suggests the nearest token matches.

    It works with any web framework (React, Vue, Svelte, etc.) as long as the app is accessible via a URL or local dev server. It can ingest design tokens from Tailwind configs, CSS variables, or standard JSON token files.

    Yes, you can define custom interaction steps to navigate through login screens, modals, or specific user flows. This allows the agent to audit styles that are hidden behind user actions rather than just checking the landing page.

    You will need Node.js 18+ and Playwright installed in your environment. The skill is designed to be compatible with major AI agent runners that support local script execution and browser automation.

    New in Frontend & Web Apps

    Free