Skip to content
SkillModel-led1,871 tokens in use

frontend-design

Design guidance that pushes Claude toward distinctive, intentional UI instead of templated defaults, from palette and type to layout and copy.

197,763 stars on the repositoryanthropics/skills · skills/frontend-design
/plugin install example-skills@anthropic-agent-skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing frontend-design costs

41tokens idle
1,871tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts and code blocks. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter11/15
  • Structure25/25
  • Examples3/15
  • Output10/15
  • Size15/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity72
  • Maintenance32
  • Community68
  • Maturity36
  • Momentum40

frontend-design is Anthropic’s agent skill for building new user interfaces or reshaping existing ones with a clear point of view. It asks Claude to work like the design lead at a studio whose client has already rejected templated proposals: ground the design in the subject matter, make deliberate choices about palette, typography and layout, and spend boldness in one place.

The skill names the traits AI-generated design currently clusters around, such as warm cream backgrounds with terracotta accents, acid-green on near-black, the SaaS card kit and all-caps eyebrow labels, and tells Claude to avoid them wherever the brief leaves an axis free. It sets a two-pass process: a compact token plan reviewed against the brief, then the build, with a quality floor for responsiveness, keyboard focus, reduced motion and accessible color, and guidance on writing interface copy.

What You Get

  • A design method that grounds palette, type and layout in the product’s own subject matter and vernacular
  • A named list of common AI-generated design tells to avoid unless the brief asks for them
  • A two-pass process: a token plan with 4-6 named hex colors, typefaces, an ASCII-wireframed layout and principles, reviewed before any code
  • Typography rules covering type scale, line length under 80 characters and headline treatments to avoid
  • A quality floor: responsive to mobile, visible keyboard focus, reduced motion respected and accessible color
  • Guidance on interface copy: active voice, consistent action names, and errors and empty states that give direction

Common Use Cases

  • Designing a landing page or hero section that should not look like every other generated page
  • Reshaping an existing UI that reads as a template into one with a specific visual identity
  • Planning a design system’s palette, type and layout before writing CSS
  • Writing button labels, toasts, error messages and empty states that match the interface’s voice

What Makes It Unique

Rather than handing Claude a style to copy, the skill works by subtraction: it lists the specific looks AI-generated design falls into, down to hex values like #F4F1EA and #D97757, the rgba(0,0,0,.1) card shadow and the arrow appended to buttons, and asks Claude to treat them as defaults, not choices. The brief’s own words always win, so a requested look is still followed. It also pairs visual design with copywriting, treating words as design content.

Under The Hood

The skill is a single SKILL.md of instructions with no scripts or reference files; its folder also holds the license text. When it loads, Claude reads the guidance and applies it in its own work: identifying the subject and audience, drafting a token plan with colors, typefaces and an ASCII wireframe, revising any part that reads as generic, then writing the code with care for CSS selector specificity. It runs no commands and calls no services itself. It suggests taking screenshots to critique the build when the environment supports it, and keeping notes on what was tried when there is somewhere to keep them.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers