Cansın Cengiz
Docs

Claude Skills for Web Design

19 Claude Code skills I use for web design: aesthetic direction, reference analysis, motion, brand visuals and testing.

Category
Tools
Updated
October 3, 2026
Sources
19

The Claude Code skills I use when designing and building websites, grouped by category with what each one does and when it comes into play.

Section 01

What a skill is and how to install it

A skill is a folder that packages instructions, scripts and resources for a specific job, with a SKILL.md file at its centre. When a task matches what the skill describes, Claude brings it in on its own.

  1. 01

    Download the skill folder from the link (the folder that contains SKILL.md).

  2. 02

    Put it in the project's .claude/skills/ directory to use it in one project, or in ~/.claude/skills/ to use it everywhere.

  3. 03

    Restart Claude Code. The skill kicks in on its own, or you can call it directly with /skill-name.

Section 02

Design direction and aesthetics

Section 03

From reference to design

  • 09View on GitHub

    design-dna

    Turning a reference design into a readable profile

    Extracts a design's DNA from images, screenshots or URLs across three dimensions: design system (tokens), style (qualitative feel) and visual effects (Canvas, WebGL, 3D, particles, scroll effects). The result is a JSON profile that new designs can be generated from.

    When to use it

    When you want to decode a site you admire and carry its language consistently into your own project.

  • 10View on GitHub

    imagegen-frontend-web

    A separate design reference image for every section

    Generates premium, conversion-aware website design references. The rule is strict: one separate horizontal image per section, so an 8-section landing page gets 8 images. It enforces composition variety, varied hero scales and one consistent palette across all images.

    When to use it

    When you want to see the page visually, section by section, before writing code.

  • 11View on GitHub

    image-to-code

    Design image first, then matching code

    For visually important web work, it first generates the design images itself, analyses them in depth, then builds the site to match them as closely as possible. Written for Codex.

    When to use it

    When you want both the design and the build in one image-to-code flow.

Section 04

Motion and interaction

  • 12View on GitHub

    scroll-craft

    Immersive, scroll-driven landing pages

    Turns scroll into the timeline: video scrubs frame by frame under the wheel, sections pin and advance, headlines assemble line by line and the page ground shifts colour as you travel. It interviews you about the brand first, picks a different skeleton and signature move for each page, and verifies the result by screenshotting its own scroll.

    When to use it

    When a site should feel like an experience rather than a document; for product and brand showcases.

  • 13View on GitHub

    algorithmic-art

    Generative art with p5.js

    Anthropic's official skill. Creates algorithmic art in p5.js with seeded randomness and interactive parameters: flow fields, particle systems, generative patterns. It makes original work rather than copying existing artists.

    When to use it

    When you need a code-generated visual for a hero background, a loading screen or a brand texture.

  • 14View on GitHub

    motion-studio

    Motion graphics and promo films rendered from code

    Builds a short film in code, either as pure motion graphics or as a motion layer on a generated take. Each film gets its own style guide, a shot list on the beat grid, kinetic type on springs, a synthesised score and sound design, and a critique loop over every rendered frame. Pairs with reel-taste.

    When to use it

    When you need a hero video, a product launch film or a short promo for social media.

Section 05

Visuals and brand

  • 15View on GitHub

    brandkit

    Brand guideline boards and logo systems

    Generates high-end brand guideline boards, logo systems, identity decks and visual-world presentations. Tuned for minimalist, cinematic, editorial, dark-tech and luxury brand worlds among others, with an emphasis on symbolic meaning and sparse typography in logo concepts.

    When to use it

    When you want to pin down a brand's identity visually before designing the site.

  • 16View on GitHub

    canvas-design

    Static visuals as PNG and PDF

    Anthropic's official skill. Starts from a design philosophy to produce posters, art pieces and other static designs as PNG and PDF, using a bundled library of openly licensed fonts.

    When to use it

    When you need a single static piece such as a cover image, an OG image or a poster.

Section 06

Build and test

  • 17View on GitHub

    webapp-testing

    Testing a local site with Playwright

    Anthropic's official skill. Opens, drives and tests local web apps with Playwright: verifies frontend behaviour, debugs UI issues, captures screenshots and reads browser logs.

    When to use it

    When checking that a change really looks and works right in the browser.

  • 18View on GitHub

    find-docs

    Up-to-date library documentation

    Fetches current documentation and code examples for any programming library or framework through Context7, including setup questions, migration guides and version-specific docs.

    When to use it

    When you need to be sure of an API in a library such as Next.js, Tailwind or framer-motion, and the model's knowledge may be out of date.

  • 19View on GitHub

    full-output-enforcement

    Complete, untruncated code output

    Overrides the model's default truncation habits: enforces complete code, bans placeholder patterns and splits output cleanly when it hits a token limit.

    When to use it

    When a long component or page has to be written out in full, in one go.