All products

Design interface components with AI

Build coherent buttons, forms and dialogs around existing tokens, with clear states and practical keyboard and responsive checks.

Design interface components with AI around what people actually do with a control. The workflow starts with the project's existing buttons, fields and tokens, then specifies variants and the states each element needs.

Ask it to improve an order form while preserving the brand palette, for example. The useful result includes loading, validation and cancellation behaviour, rather than another isolated button style. It works in your current stack and calls for a real browser review before treating the interface as finished.

What it does

  • Specify default, focus, pressed and disabled states
  • Keep loading controls stable and avoid duplicate activation
  • Replace placeholder-only fields with visible labels
  • Give empty collections a meaningful next action
  • Reuse project primitives instead of adding competing variants
  • Review menus, tabs and dialogs with keyboard controls

Install and use

  1. Provide the selected folder and its design-taste, spacing, typography and color companions to a file-capable agent. Supply the authorised project and a browser for verification.
  2. Identify the control, expected behaviour and permitted changes. Preserve existing primitives, brand decisions and token sources.
  3. Review the component inventory and state matrix before implementing agreed changes in one shared location.
  4. Exercise the result with a keyboard, narrow viewport and enlarged text. Check actual loading and error states; record anything not tested.

FAQ

Source

Made by Imran Ahmed. Original skill: components, published on GitHub under the MIT licence.

View on GitHub