Modern Css logo

Modern Css

Organization
moderncss
modern-css

Apply these opinionated rules whenever CSS is written, refactored, or reviewed, so agents and people write it consistently. They sit on top of the Modern Web Guidance skill and settle the choices it leaves open across architecture, layout, typography, colors, and motion. Use this skill when the user asks things like "set up the base styles for this project", "style this component", "make this responsive", "add dark mode", "scope these styles", "fluid type scale", or "refactor my styles to modern CSS", when they ask to build a page or component, and for any other CSS task, even when they don't say "CSS".

Overview

Publishermoderncss
Repositoryskills
Skill namemodern-css
Stars
5
Forks
0
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by moderncss on GitHub. Read the source before you install it.

Installation

Install the Modern Css AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/moderncss/skills/tree/main/skills/modern-css
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/moderncss/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/skills/modern-css .claude/skills/modern-css
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Modern Css in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Modern Css on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Modern Css is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Modern CSS

Opinionated rules for modern CSS, so agents and people write it consistently.

These rules sit on top of the Modern Web Guidance skill, which must be installed. It explains what each feature does, when it applies, and how to fall back. It leaves choices open: which color space, whether to scope, when to nest, how to size type. These rules settle them; retrieve its guides for everything else. Where the two disagree, the rule wins.

The stylelint-config-modern package enforces every rule a linter can. This skill gets agents writing those from the start, and carries the rest.

Browser support policy: features within Baseline Newly Available are used natively, without @supports fallbacks, as progressive enhancement.

Rules

Architecture

Organizing styles (@layer)
  • Rule: Put every style rule in a @layer.
  • Constraint: Avoid unlayered element defaults and resets.
  • Rationale: An unlayered rule outranks every layer; with everything layered, layer order alone decides conflicts.
  • References: @layer on MDN.
  • Example:
css
@layer elements, components;

@layer elements {
  a {
    text-decoration-skip-ink: auto;
  }
}
Encapsulating styles (@scope)
  • Rule: Use @scope for every component's styles, with a limit (@scope (outer) to (inner)) where the component hosts content it doesn't own: embedded components, slot content, or user content.
  • Constraint: Avoid component styles outside a scope, and an unlimited scope over content the component doesn't own.
  • Rationale: Scoped selectors can't bleed into other components, and the limit stops the outer scope's styles reaching projected content. Sometimes called donut scoping.
  • References: @scope on MDN.
  • Example:
css
@scope (.card) to (.content) {
  h2 {
    font-size: var(--large);
  }
}
Nesting rules and at-rules (&)
  • Rule: Nest rules and at-rules inside their parent rule; write & where the nested selector attaches to the parent (&:hover) and omit it for combinators (b, not & b).
  • Constraint: Avoid unnested selectors (e.g. a {} a:hover {}).
  • Rationale: Keeps a rule's states and queries beside it in one block, and stylelint-config-modern enforces the implicit & form.
  • References: & nesting selector on MDN.
  • Example:
css
a {
  color: var(--blue);

  &:hover {
    text-decoration: underline;
  }

  code {
    font-variant-numeric: tabular-nums;
  }
}
Ranged queries (20em < width <= 40em)
  • Rule: Use range syntax for size queries (e.g. @container (20em < width <= 40em)), with ranges that don't overlap.
  • Constraint: Avoid min-width/max-width and overlapping ranges (e.g. width <= 20em alongside width >= 20em, which both match at 20em).
  • Rationale: Each size lands in exactly one condition, so no condition has to override another.
  • References: Media query range syntax on MDN.
  • Example:
css
.card {
  @container (width <= 20em) {
    grid-template-columns: 1fr;
  }

  @container (20em < width <= 40em) {
    grid-template-columns: 1fr 1fr;
  }

  @container (width > 40em) {
    grid-template-columns: repeat(3, 1fr);
  }
}

Layout

Flow-relative layout (*-inline-*, *-block-*, cqi/vi, start/end)
  • Rule: Use flow-relative properties (e.g. padding-block-start, inset-inline, inline-size), units (e.g. cqi, cqb, vi), and keywords (e.g. text-align: start) for layout.
  • Constraint: Avoid physical equivalents (e.g. padding-top, width, cqw, vw, text-align: left), even where the layout would never flip.
  • Rationale: Keeps the box model on the axes flexbox and grid already use, and stylelint-config-modern enforces it.
  • References: Logical properties and values on MDN.
  • Example:
css
.card {
  max-inline-size: 60ch;
  padding-block-start: 2cqi;
  text-align: start;
}
Two-value display (display: block flex)
  • Rule: Use the two-value display syntax (e.g. display: block flex, display: inline grid).
  • Constraint: Avoid the single-value keywords (e.g. display: block, display: flex, display: inline-block).
  • Rationale: Separates how the box sits in its parent from how its children lay out, so inline flex and inline flow-root replace the special-case keywords inline-flex and inline-block.
  • References: Multi-keyword display syntax on MDN.
  • Example:
css
nav {
  display: block flex;
}

.card {
  display: block grid;
}
Fluid spacing (cqi)
  • Rule: Size spacing with container units (e.g. padding: 2cqi), clamped where it needs a floor or ceiling.
  • Constraint: Avoid fixed spacing (e.g. padding-block: 16px, margin-inline: 1rem).
  • Rationale: Spacing scales with the container, so components adapt wherever they're placed, without the seams breakpoints leave.
  • References: Container query length units on MDN, Responsive design: seams & edges..
  • Example:
css
.card {
  container: card / inline-size;
  gap: 2cqi;
  padding: clamp(1rem, 0.5rem + 2cqi, 2rem);
}
Auto-growing text areas (field-sizing: content)
  • Rule: Use field-sizing: content on text areas, with an explicit inline-size and min-block-size/max-block-size in lh.
  • Constraint: Avoid fixed block sizes (e.g. block-size: 8rem) and field-sizing: content on inputs, which shrink to their value's width.
  • Rationale: Typed text stays in view instead of scrolling inside a fixed box, and lh bounds follow the line height as the type scales.
  • References: field-sizing on MDN.
  • Example:
css
textarea {
  field-sizing: content;
  inline-size: 100%;
  max-block-size: 12lh;
  min-block-size: 3lh;
}

Typography

Fluid type sizes (clamp())
  • Rule: Use clamp() with container units for font sizes, on a type scale whose ratio grows with the container, e.g. Major Second (1.125) on narrow containers and Major Third (1.25) on wide ones.
  • Constraint: Avoid fixed font sizes (e.g. px, rem) and central values without a rem term (e.g. clamp(1.75rem, 5cqi, 2.25rem)).
  • Rationale: Sizes text to its container rather than at breakpoints, and the rem term keeps zoom working.
  • References: clamp() on MDN, Designing with fluid type scales.
  • Example:
css
:root {
  --medium: clamp(1.3125rem, 1.1821rem + 0.6522cqi, 1.6875rem);
  --large: clamp(1.75rem, 1.5761rem + 0.8696cqi, 2.25rem);
}
Widow and orphan words (text-wrap)
  • Rule: Use text-wrap: balance on headings and text-wrap: pretty on all other text.
  • Constraint: Avoid default wrapping outside of inputs and text areas.
  • Rationale: Removes widows and orphans by default rather than where someone remembered to opt in.
  • References: text-wrap on MDN.
  • Example:
css
h1,
h2,
h3 {
  text-wrap: balance;
}

body {
  text-wrap: pretty;
}

Colors

Perceptual uniform lightness (oklch())
  • Rule: Use oklch() for all colors.
  • Constraint: Avoid hex, rgb(), hsl() and other color formats.
  • Rationale: Lightness is perceptually uniform, so text contrast holds across hues.
  • References: oklch() on MDN.
  • Example:
css
:root {
  --success: oklch(40% 0.15 150deg);
  --danger: oklch(40% 0.2 25deg);
}
Respecting color preferences (light-dark())
  • Rule: Use light-dark() for every color that differs between light and dark schemes.
  • Constraint: Avoid prefers-color-scheme blocks that redeclare colors, and colors that don't adapt.
  • Rationale: Each token carries both values in one place, so no scheme can be missed when a color changes.
  • References: light-dark() on MDN.
  • Example:
css
:root {
  --surface: light-dark(oklch(98% 0 0deg), oklch(16% 0 0deg));
  --text: light-dark(oklch(35% 0 0deg), oklch(75% 0 0deg));
}
Relative color functions (oklch(from /* .. */) & color-mix() )
  • Rule: Derive related colors with relative color syntax (e.g. oklch(from var(--primary) l c calc(h - 10deg))) or color-mix() in oklab.
  • Constraint: Avoid hardcoding colors that relate to other colors, and lightness-only moves on saturated colors (e.g. oklch(from var(--primary) 90% c h)).
  • Rationale: Related colors follow their source when it changes.
  • References: relative colors on MDN, color-mix() on MDN.
  • Example:
css
:root {
  --primary: oklch(55% 0.25 350deg);
  --primary-hover: oklch(from var(--primary) l c calc(h - 10deg));
  --primary-tint: color-mix(in oklab, var(--primary), oklch(100% 0 0deg) 80%);
}

Motion

Respecting motion preferences (prefers-reduced-motion)
  • Rule: Use prefers-reduced-motion: no-preference when applying large animations and transitions.
  • Constraint: Avoid prefers-reduced-motion: reduce.
  • Rationale: Motion becomes opt-in, so no animation needs a reduced-motion override. With reduce, every animation needs one, and it's easy to miss one.
  • References: prefers-reduced-motion on MDN.
  • Example:
css
@media (prefers-reduced-motion: no-preference) {
  .hero {
    animation: bounce-in 0.5s ease;
  }
}

Applying these rules

When asked to author, refactor, or review CSS:

  1. Identify which rules apply to the task; where several do, apply them all.
  2. Read surrounding files to see which rules are already in use; match their patterns.
  3. Apply the rules consistently across the change set, not just at the touch points.
  4. Retrieve the Modern Web Guidance guides for the task and apply them within these rules, skipping their fallbacks under the browser support policy above.
  5. If the project runs Stylelint with stylelint-config-modern, run it on the changed files and fix what it reports before finishing.

Gotchas

Traps neither the rules above nor Modern Web Guidance surface:

  • clamp() central values need a rem term: clamp(1.75rem, 1.5761rem + 0.8696cqi, 2.25rem), not the clamp(1rem, 5cqi, 2.5rem) form Modern Web Guidance's fluid-scaling guide shows.
  • grid-template-rows: subgrid only inherits tracks the child claims. Pair it with grid-row: span N; without the span the child gets one row and aligns with nothing. Modern Web Guidance's css-layout guide shows the span without saying why.
  • container-type already applies containment: a component with container-type: inline-size has layout, style and inline-size containment, so it needs no contain. The contain: layout style paint that Modern Web Guidance's performance guide puts on widgets adds paint containment, which clips overflowing descendants like overflow: clip.

Examples

This site follows the rules:

Fetch one only when the project has no CSS of its own to match.

Frequently asked questions

What does the Modern Css AI skill do?

Apply these opinionated rules whenever CSS is written, refactored, or reviewed, so agents and people write it consistently. They sit on top of the Modern Web Guidance skill and settle the choices it leaves open across architecture, layout, typography, colors, and motion. Use this skill when the user asks things like "set up the base styles for this project", "style this component", "make this responsive", "add dark mode", "scope these styles", "fluid type scale", or "refactor my styles to modern CSS", when they ask to build a page or component, and for any other CSS task, even when they don...

Why use Modern Css on TypingMind?

Because you install it once and use it with any model. Modern Css is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Modern Css in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/moderncss/skills/tree/main/skills/modern-css. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Modern Css?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Modern Css?

As many as you like. As long as a model supports skills, you can use Modern Css with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Modern Css AI skill free?

Yes. It is published on GitHub by moderncss under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇