agentleFS
coolxeocoolxeo/grok-bot-skills/skills/design/design-taste-frontend/SKILL.md

design-taste-frontend

Skill8 starsChanged 2 days agoMITView on GitHub

Use when building a landing page, portfolio, or marketing site from a brief, so the result looks chosen for this subject instead of like a template. Reads the brief, plans tokens, builds, then checks the rendered page.

190 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use stitch-design-taste, design-taste-frontend-v1 and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • HK-hubSkill

    stitch-design-taste

    Semantic Design System Skill for Google Stitch. Generates agent-friendly DESIGN.md files that enforce premium, anti-generic UI standards — strict typography, calibrated color, asymmetric layouts, perpetual micro-motion, and hardware-accelerated performance.

    Also in 59 projects that use this

  • HK-hubSkill

    design-taste-frontend-v1

    The original v1 taste-skill, preserved for projects depending on its exact behavior. The current default is `design-taste-frontend` (v2 experimental), which is a substantial rewrite. Use this v1 install name only if you need exact backward compatibility.

    Also in 29 projects that use this

  • UtkarshDubeyGITSkill

    firecrawl-website-design-clone

    Extract any website's design system into an agent-ready DESIGN.md using Firecrawl scrape evidence. Use when the user wants colors, fonts, spacing, components, layout patterns, or brand/UI guidance from a website so AI agents can create new websites, clone a look, or build pages inspired by that design.

    Also in 3 projects that use this

  • hjdspaceSkill

    flutter-setup-declarative-routing

    Configure `MaterialApp.router` using a package like `go_router` for advanced URL-based navigation. Use when developing web applications or mobile apps that require specific deep linking and browser history support.

    Also in 3 projects that use this

  • hjdspaceSkill

    flutter-setup-localization

    Add `flutter_localizations` and `intl` dependencies, enable "generate true" in `pubspec.yaml`, and create an `l10n.yaml` configuration file. Use when initializing localization support for a new Flutter project.

    Also in 3 projects that use this

  • LangChainSkill

    embedded-captions

    Add captions or subtitles to an existing single-subject talking-head video without editing the footage. Use for plain verbatim captions, cinematic captions embedded behind the subject, VFX captions, “炸/特效/酷炫字幕,” or a named identity from the 35-style catalog. Route by visual identity, not by backend engine. The quiet `anchor` rail is the default; embed every word only when the user explicitly wants a fully cinematic treatment. The workflow runs locally end to end, including transcription and subject matting; split multi-shot footage before applying it.

    Also in 5 projects that use this

  • +3 more from hjdspace

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 186 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write a compact plan and check it against the brief:

    line 20
  2. Distinctive:

    Check that every dependency exists in the project before importing it.

    line 43

“Common” is measured: kept by at least 6 in 10 of the 186 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Design taste: frontend

Build a page that looks like someone made choices for this subject and this audience. Most generated pages fail by reaching for the same defaults every time; this skill exists to stop that.

1. Read the brief

Before any code, write a one-line design read: what the page is for, who it's for, the one action it should drive, and the mood in three words. Pull it from the brief, linked references, screenshots, existing brand assets, and the audience. If a key piece is missing, ask one question; otherwise state your assumptions and go.

Constraints win over taste: accessibility needs, regulated industries, kids, public sector, or an existing brand system. If the brand has a design system or official component package, use it.

2. Plan before building

Write a compact plan and check it against the brief:

  • Color: 4-6 named hex values with roles (background, text, accent, muted, border).
  • Type: one or two families, chosen for this subject, with a clear size scale. Body lines under about 75 characters.
  • Layout: a one-line idea per section plus a rough ASCII wireframe of the hero.
  • The one bold thing: the single element people will remember. Everything else stays quiet.

Then ask: would I have produced this plan for any similar page? If yes, change that part and say why.

3. Defaults to avoid (unless the brief asks)

  • Warm cream background with a serif headline and a terracotta accent.
  • Near-black page with one neon accent; purple-blue gradients; glow.
  • Identical rounded cards with the same soft shadow, used for everything.
  • An all-caps eyebrow over every heading, 01 / 02 / 03 markers on content that isn't a sequence, one word in the headline highlighted in a different color.
  • Centered hero, three feature cards, logo strip, testimonial carousel, in that order, for every product.
  • Fake product screenshots built from divs, stock "Jane Doe" testimonials, invented logos or metrics.
  • Fade-and-slide on every section. Use one deliberate motion moment, or none.

4. Build

  • Real copy written for the end user: plain verbs, sentence case, buttons that say what happens ("Start free trial", not "Submit"). Run headlines through no-ai-slop. No em dashes in UI copy.
  • Real images if they exist. If they don't, leave clearly labeled slots and list them at the end instead of faking them.
  • Check that every dependency exists in the project before importing it.
  • Quality floor, no announcement needed: responsive to 360px, visible keyboard focus, prefers-reduced-motion respected, text contrast at WCAG AA, links are links and buttons are buttons, no transition: all.

5. Look at it

Open the page in a browser at mobile and desktop widths and take screenshots. Check the hero above the fold, spacing rhythm, contrast, and that the bold thing reads as intended. Fix what looks off, then hand over the screenshots with the code.

Inspired by

  • Leon Lin, Taste Skill (MIT). Does well: reads the brief before choosing a look, a one-line design read, a long list of generated-design tells, and a ban on fake screenshots. Missing: about 1,200 lines to load, it rules out dashboards and product UI, and it makes dark mode mandatory for consumer pages.
  • Anthropic, frontend-design skill (Apache 2.0). Does well: plan a small token system, review it against the brief before building, and spend boldness in one place. Missing: no step to open the rendered page and look at it.
  • Vercel, Web Interface Guidelines (MIT). Does well: a concrete checklist for keyboard, focus, motion, forms, and layout. Missing: no aesthetic direction at all.

What this version adds

  • A short, loadable skill: design read, token plan, defaults to avoid, build, look.
  • A browser screenshot check at two widths before handing over.
  • Copy rules tied to no-ai-slop, and an accessibility floor drawn from interface guidelines.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.