All skills
google-labs-code avatar

/extract-design-md

@681a283 official

Extract a comprehensive design system (DESIGN.md) directly from frontend source code — React, Vue, Svelte, Angular, plain HTML/CSS, or any web framework. Analyzes component files, stylesheets, Tailwind configs, theme definitions, and design tokens to produce a rich, Stitch-compatible design system document. Use this skill whenever the user wants to reverse-engineer a design system from an existing codebase, audit the visual language of a project, extract design tokens from source files, or understand the styling patterns in a frontend repo — even if they just say "what does this app look like?" or "pull out the design from this code."

  • 7 files
  • 39.7 KB
  • Updated 5 months ago
  • GitHub

Use this Skill: https://skilld.dev/gh/google-labs-code/stitch-skills/extract-design-md

This session only. Nothing lands on disk.

examplesDESIGN.md

≈1.7k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Brand & Style

The brand personality is high-performance, sophisticated, and dependable. It is designed for the modern skier who demands clarity and precision in harsh environments. The visual language evokes the crystalline beauty of a mountain peak while prioritizing the rugged utility required for outdoor navigation.

The design system utilizes Glassmorphism to achieve a "frozen" aesthetic, mimicking the properties of ice and packed snow. This is paired with a High-Contrast philosophy to ensure that critical data—like trail status and weather alerts—remains legible under intense mountain sunlight or the low-visibility conditions of a snowstorm.

Colors

The palette is anchored by Deep Peak Blue, a dark navy that provides a grounding contrast for white text and maps. Powder White serves as the primary canvas, ensuring the interface feels airy and cold.

Safety Orange is reserved strictly for primary calls to action, hazard warnings, and emergency trail closures, ensuring immediate eye-tracking. Electric Blue (the info status) is used for weather updates and general interactivity. For trail difficulty levels, standard industry colors (Green, Blue, Black) are used but rendered with high-saturation values to pop against the white backgrounds.

Typography

The typography system uses Inter for core functional text and body copy due to its exceptional x-height and legibility at small sizes. Lexend is introduced for labels and statistics; its hyper-readable, athletic character suits the "active" nature of skiing data like speed and vertical drop.

To combat outdoor glare, font weights are generally heavier than standard web applications. Display styles use tight tracking and heavy weights to create a sense of strength. Labels always utilize high-contrast coloring against their backgrounds.

Layout & Spacing

This design system uses a Fluid Grid model with generous safe areas. Elements are spaced using a strict 4px/8px baseline rhythm to maintain a disciplined, professional appearance.

Touch targets are intentionally oversized (minimum 48x48px) to accommodate users wearing gloves. Padding within cards and containers is kept spacious (lg or xl) to prevent the interface from feeling cluttered when displaying complex trail data.

Elevation & Depth

Depth is conveyed through Glassmorphism and backdrop filters rather than traditional heavy shadows. Surfaces use a hierarchy of transparency:

  1. Base Layer: Solid 'Powder White' or 'Deep Peak Blue'.
  2. Middle Layer (Cards/Modals): 70% opacity white with a 20px backdrop blur and a thin 1px 'Ice' border (white at 40% opacity).
  3. Top Layer (Floating Actions): 90% opacity with a subtle 4px ambient shadow to separate the element from the blurred background.

This creates a "stacked ice" effect that maintains visual clarity while giving the UI a premium, modern feel.

Shapes

The shape language is defined by modern, organic rounded corners. This softens the high-contrast color palette and makes the app feel approachable.

Interactive elements like buttons and chips utilize rounded-lg (1rem) or rounded-xl (1.5rem) to suggest a tactile, "pebble" feel. Large containers like map overlays use 1rem corners, while small status indicators for trail difficulty may use pill shapes for instant recognition.

Components

Buttons

Primary buttons use a solid 'Safety Orange' or 'Deep Peak Blue' fill with white text. High-contrast outlines are used for secondary actions. All buttons must feature a minimum height of 52px for gloved-hand accessibility.

Cards & Modals

Cards utilize the glassmorphism effect—semi-transparent backgrounds with a 1px white border. This ensures that map imagery or photos underneath remain slightly visible, maintaining the "frozen" aesthetic.

Trail Status Chips

Pill-shaped indicators with high-saturation icons. Green (Easy), Blue (Intermediate), Black (Expert), and Double Black (Extreme). These must include text labels alongside icons to ensure accessibility for colorblind users.

Input Fields

Fields feature a 'Cloud Gray' background with a thick bottom border that transforms into 'Electric Blue' on focus. Labels are always visible above the field to ensure the user doesn't lose context in bright environments.

Weather Widgets

Bold, thick-stroke iconography (minimum 2pt stroke) to represent sun, snow, or wind. These icons should be large and accompanied by 'Lexend' typography for temperature and wind speed.

Source: SKILL.md on GitHub

No third-party reports yet.

Signed by skilld at 681a283. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Steadyupdated 5 months ago
What it can do
Runs commands Reads files Edits files
All 5 allowed tools
stitch*:*BashReadWriteweb_fetch

README badge

README badge for google-labs-code/stitch-skills/extract-design-md

Analyzes frontend source code (React, Vue, Svelte, Angular, plain CSS) to extract a comprehensive design system document (DESIGN.md) that captures colors, typography, spacing, and component patterns without building or rendering the app. Use this when you need to audit styling, document visual language, or understand design tokens directly from source files.

Generated from the current SKILL.md.

Does this skill work without building or running the app?
Yes. It reads source files directly — stylesheets, component files, theme configs, Tailwind setups — so you don't need dependencies installed or the app running.
What frameworks does this skill support?
React, Next.js, Vue, Nuxt, Svelte, SvelteKit, Angular, and plain CSS/SASS/Less. It also handles CSS-in-JS libraries like styled-components and Emotion, plus Tailwind and design token files.
What does the output look like?
A DESIGN.md file placed at `.stitch/DESIGN.md` with YAML frontmatter and sections covering visual theme, color palette, typography, component styles, layout principles, and Stitch generation notes.
Can this extract colors, fonts, and spacing from Tailwind config?
Yes. The skill reads tailwind.config.js/ts to extract custom colors, fonts, and spacing scales, along with scanning component files and stylesheets.
Is this skill different from the design-md skill?
Yes. The design-md skill analyzes rendered HTML. This skill works from source code only, making it faster and useful when the app can't be built or run.

Generated from the current SKILL.md. These answers refresh after source changes.