atomic-design
Skill4 starsChanged 8 months agoCC0-1.0View on GitHub
Brad Frost's Atomic Design methodology for UI component hierarchies: atoms (indivisible elements), molecules (small groups of atoms), organisms (complex sections of molecules), and templates (page layouts). Enforces bottom-up composition (never skip levels), presentational components (data via props, events via callbacks, no data fetching), design tokens for all visual properties, and composition over inheritance. Use when building user interfaces, creating component libraries, organizing frontend code, designing form systems, or structuring any UI. Triggers on: "build a component", "component hierarchy", "design tokens", "presentational components", "composition over inheritance", "React/Vue/SwiftUI components", "atomic design". Applies to any UI framework.
We split this skill into the rules it gives an agent
14 rules. About 0 are common, 14 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈14
Distinctive · estimate
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
- Distinctive: line 43
Start with the smallest reusable elements and compose upward.
- Distinctive: line 72
Start with atoms when building new UI
- Distinctive: line 74
Keep atoms under 50 lines, molecules under 100
- Distinctive: line 78
Create a molecule that does not compose atoms from your system
- Distinctive: line 100
Put API calls inside atoms, molecules, or organisms
- Distinctive: line 110
Every color, spacing value, and font size in a component must come from a token
- Distinctive: line 126
Create one-off token names for single components
- Distinctive: line 135
Keep the component tree flat -- prefer siblings over deep nesting
- Distinctive: line 136
When you need a variant, compose a new molecule from atoms rather than adding flags to an existing molecule
- Distinctive: line 171
NOT presentational. The test: could this component render identically in a Storybook story with mock props?
- Distinctive: line 177
After completing work guided by this skill, verify:
- Distinctive: line 185
If any criterion is not met, revisit the relevant practice before proceeding.
- Distinctive: line 194
domain-modeling: Read models from the domain define what data components receive as props.
- Distinctive: line 196
tdd: Test components in isolation at each level -- atom tests, molecule tests, organism tests.
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Atomic Design
Value: Simplicity and communication. Building UI from small, named, composable pieces makes the interface understandable to everyone on the team and prevents the complexity that comes from monolithic components.
Purpose
Teaches how to organize UI components into a hierarchy of increasing complexity: atoms, molecules, organisms, and templates. Each level has clear responsibilities and composition rules. The outcome is a component system where every piece is reusable, testable in isolation, and named in a shared vocabulary.
Practices
Build Bottom-Up Through Four Levels
Start with the smallest reusable elements and compose upward. Never skip a level.
The four levels:
- Atoms: Indivisible UI elements. A button, an input, a label, an icon. One visual element, one responsibility. Atoms reference design tokens for all visual properties (color, spacing, typography).
- Molecules: Small groups of atoms functioning as a unit. A form field (label + input + error message). A search bar (input + button + icon). One interaction pattern per molecule.
- Organisms: Complex components composed of molecules and atoms that form a distinct section of the interface. A navigation header, a complete form, a data table. One feature area per organism.
- Templates: Page-level layouts that arrange organisms into a complete view. A dashboard template, a list-detail template. Templates define structure and content slots, not specific data.
Example:
Atom: Button, Input, Label, ErrorMessage
Molecule: FormField (Label + Input + ErrorMessage)
Organism: LoginForm (FormField + FormField + Button)
Template: AuthPage (Header + LoginForm + Footer)Do:
- Start with atoms when building new UI
- Name components by what they ARE, not what data they show
- Keep atoms under 50 lines, molecules under 100
Do not:
- Build organisms directly from raw markup -- extract atoms first
- Create a molecule that does not compose atoms from your system
- Skip to templates before organisms exist
Keep Components Presentational
Components render UI. They receive data as props. They do not fetch data, manage business logic, or hold application state.
- Pass all data through props or equivalent
- Emit events for user actions -- do not handle side effects
- Separate data containers from presentational components
Example:
Presentational (good):
UserCard({ name, email, avatar }) -> renders UI
Container (separate):
UserCardContainer() -> fetches data, passes to UserCardDo not:
- Put API calls inside atoms, molecules, or organisms
- Couple a component to a specific data source
- Mix rendering logic with business logic in the same component
Use Design Tokens for All Visual Properties
Extract every design decision (colors, spacing, typography, shadows, radii) into named tokens. Components reference tokens, never raw values.
- Define tokens as the first step of any new design system
- Every color, spacing value, and font size in a component must come from a token
- Changing a token updates every component that references it
Example:
/* Tokens */
--color-primary: #0066cc;
--spacing-sm: 8px;
--spacing-md: 16px;
/* Component uses tokens, not values */
.button { background: var(--color-primary); padding: var(--spacing-sm); }Do not:
- Hard-code
#0066ccor8pxin any component - Create one-off token names for single components
- Define tokens that are never used (tokens should earn their place)
Compose, Do Not Inherit
Build complex components by nesting simpler ones. Do not extend base components through class inheritance or deep prop-forwarding chains.
- Pass children or slots to compose layout
- Keep the component tree flat -- prefer siblings over deep nesting
- When you need a variant, compose a new molecule from atoms rather than adding flags to an existing molecule
Do:
IconButton = Icon + Button(composition)Card > CardHeader + CardBody(slots)
Do not:
FancyButton extends Button(inheritance)- A single Button component with 15 variant props
Enforcement Note
Advisory in all modes. Component hierarchy and token discipline are self-enforced.
Hard constraints:
- Token-only references (no raw values in components):
[RP]
Constraints
- "Never skip a level": An atom that's actually a molecule (it composes multiple visual elements) is skipping a level even if you name it "atom." The classification is based on what the component IS, not what directory it's in. If your "atom" has 3 internal elements with layout logic, it's a molecule.
- "No raw values": Defining a token for every unique value and then
never reusing those tokens defeats the purpose. Tokens exist for reuse
and consistency. If a token is used exactly once, ask: should this value
be shared with other components? If yes, the token is correct. If no,
the value should probably come from a more general token (e.g., use
spacing-mdnotcard-header-padding-top). - Presentational boundary: Presentational means: data in via props, events out via callbacks. Filtering data for display IS presentational (it's a view concern). Fetching data, mutating state, or calling APIs is NOT presentational. The test: could this component render identically in a Storybook story with mock props? If it needs a running backend, it's not presentational.
Verification
After completing work guided by this skill, verify:
- Every UI element traces to an atom (no raw markup in organisms/templates)
- Design tokens exist and components reference them (no hard-coded values)
- Each component has a single responsibility appropriate to its level
- Components are presentational (data passed in, events emitted out)
- The hierarchy is documented or self-evident from directory structure
If any criterion is not met, revisit the relevant practice before proceeding.
Dependencies
This skill works standalone. For enhanced workflows, it integrates with:
- design-system: The design system specification provides the token definitions, component catalog, and hierarchy that this skill implements in code.
- domain-modeling: Read models from the domain define what data components receive as props.
- tdd: Test components in isolation at each level -- atom tests, molecule tests, organism tests.
- event-modeling: Wireframes from event modeling sessions identify which components are needed.
Missing a dependency? Install with:
npx skills add jwilger/agent-skills --skill tddOther versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- Yesterday-AI/ytstack★ 1same name
- CorentinGC/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository21 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- bumpSkill
- agent-coordinationSkill
- architecture-decisionsSkill
- bootstrapSkill
- ci-integrationSkill
- code-reviewSkill
- debugging-protocolSkill
- design-systemSkill
- domain-modelingSkill
- ensemble-teamSkill
- event-modelingSkill
- factory-reviewSkill
- memory-protocolSkill
- mutation-testingSkill
- pipelineSkill
- session-reflectionSkill
- task-managementSkill
- tddSkill
- ticket-triageSkill
- user-input-protocolSkill
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
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.

