State Color Matrix

Define how every control looks in each interaction state, once, instead of screen by screen.

moderate Design Systems

What it adds

A matrix pairing each interactive component with its default, hover, active, focus, selected, disabled, and loading treatment.

What your agent is told to do

1
Enumerate the interactive components — buttons, links, rows, cards, tabs, menu items, form controls — and fill in every state for each, so that the combinations nobody thought about are visibly blank rather than silently absent.
2
Check contrast in every state, not only the default one. Hover and disabled treatments are where text most often drops below threshold, and disabled is where teams most often assume the rules no longer apply.
3
Treat hover as an enhancement that touch users will never see. Any meaning carried only by a hover treatment — that a row is actionable, that an item can be removed — must also be carried by something permanent.
4
Make selected, active, and disabled distinguishable from one another by more than colour: a selected row and a disabled row rendered as two shades of the same grey are indistinguishable to a large number of users and identical in a forced-colours mode.
5
Do not change an element's box dimensions on state change. Adding a border on hover or thickening one on selection reflows the row underneath; keep the border present and change only its colour, or draw it inside the element's own box.

Edge cases it handles

Contrast must hold in every state, including hover, selected, and disabled — a disabled label is still read, and a low-contrast hover makes text briefly unreadable at the exact moment someone is aiming at it.
Hover cannot carry meaning on touch devices, where it either never fires or sticks after a tap, so anything discoverable only on hover is undiscoverable to most mobile users.
Selected, active, and disabled must be distinguishable from each other and from focus without relying on hue alone, since colour-vision differences and forced-colours modes collapse them together.
State changes must not alter the element's dimensions, or the layout shifts under the pointer and the row moves out from under the click.
States combine: a selected row that is also hovered and also focused needs one defined result, not three treatments layered until nothing is legible.
Focus ring geometry, offset, and clipping belong to the Focus Ring System; this matrix owns only the colours the focus state uses, and must not specify a second ring.
A loading state on a control must remain visibly the same control at the same size, so the button does not shrink to a spinner and take the layout with it.

Definition of done

Every interactive component has a defined treatment for default, hover, active, focus, selected, disabled, and loading.
Contrast requirements are met in every state, including disabled.
No affordance or meaning is conveyed by hover alone.
Selected, active, focus, and disabled are distinguishable without relying on colour alone.
No state change alters an element's dimensions or shifts surrounding layout.
Combined states have a single defined result.
Focus ring geometry is delegated to the Focus Ring System rather than redefined.
The feature matches the existing design system.
No existing functionality is broken.

Related features

Z-Index Registry

Spacing Scale

Shadow Scale

How it works

1

Copy the link

Grab the Markdown instruction URL for this feature.
2

Give it to your AI

Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.
3

It inspects, then implements

Your agent reads your existing app first, then adds the feature to fit it.

Works with your stack

These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.

Need it tighter than that? Customize the feature and tell it exactly what you're running.