clarify
Decomposes underspecified design requests by classifying gaps (missing brand context, visual ambiguity, false premises, scope…
Produces self-contained HTML/CSS mockups and wireframes from design tokens and page specs — the output is always a viewable .html file, not guidance or code. Use when the request is to see a design, make a mock, build or show a prototype, create a wireframe, render a preview, or
$ npx -y skills add ryanthedev/design-for-ai --skill prototype --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/prototypeContext preview
The summary Claude sees to decide when to auto-load this skill.
Produces self-contained HTML/CSS mockups and wireframes from design tokens and page specs — the output is always a viewable .html file, not guidance or code. Use when the request is to see a design, make a mock, build or show a prototype, create a wireframe, render a preview, or
name: prototype description: "Produces self-contained HTML/CSS mockups and wireframes from design tokens and page specs — the output is always a viewable .html file, not guidance or code. Use when the request is to see a design, make a mock, build or show a prototype, create a wireframe, render a preview, or visualize what a page looks like before writing production code. Reads DESIGN.md tokens and JOURNEY.md page specs when present; uses greyscale wireframe defaults when absent. Renders via the browser MCP and returns a screenshot path when connected; emits the .html with an open note when not. Not for: choosing or defining fonts, colors, visual identity, or design tokens (use core design); critiquing, auditing, or reviewing an existing design (use core audit); planning IA, flows, or page specs (use journey); building a design token system or component library (use design-systems)." user-invocable: true argument-hint: "[page name or spec, fidelity: wireframe|styled]"
Prototype turns design tokens and page specs into a viewable artifact — a single self-contained `.html` file that opens in any browser with no build step.
Collect three inputs:
| Input | Source | If absent | |-------|--------|-----------| | **Tokens** | DESIGN.md at the project root (the CSS custom-property block from `palette.mjs`) | Wireframe mode: greyscale defaults | | **Page spec** | JOURNEY.md page spec section, or user description | Scaffold a generic page skeleton | | **Fidelity** | User request; "wireframe" = structure only, "styled" = full tokens | Default to styled when DESIGN.md present, wireframe when not |
If DESIGN.md is absent, note it and proceed in wireframe mode — emit the `.html` with a comment: `<!-- No DESIGN.md found — wireframe mode: greyscale defaults. Run /design-for-ai to establish tokens. -->`.
Load `references/mock-recipe.md`. Follow the recipe for the chosen fidelity:
Both fidelities: single file, all CSS in `<style>`, `prefers-reduced-motion` media query suppressing transitions, touch targets ≥44×44px on mobile surfaces, semantic HTML.
Write the file to `<project-root>/mocks/<page-name>.html` (create `mocks/` if needed), or to `./mock.html` when no project root is clear.
**When the browser MCP is connected:**
1. `browser_connect(mode: "launch", channel: "chrome", headless: true)` — or reuse existing connection. 2. `browser_navigate(url: "file://<absolute-path-to-html>", allow_internal: true)` — local file render; `allow_internal` is required. 3. `browser_screenshot()` — returns `{path}` (a PNG). 4. Dispatch a Haiku subagent to read the PNG and return a text summary (payload discipline: large artifact stays out of main context). 5. Return: the `.html` path, the screenshot path, and the Haiku summary.
**When the browser MCP is not connected (graceful fallback):**
Emit the `.html` path and note: "Open `<path>` in your browser to view — browser MCP not connected so no screenshot was captured." No error. The mock is still the deliverable.
Detection: if `browser_connect` fails, fall to the fallback. Never block on MCP availability.
Suggest running `/design-for-ai:mock` (which dispatches the `design-review-agent`) to critique the rendered result, or `build` for a full gated review.
| Reference file | Load when | |----------------|-----------| | `references/mock-recipe.md` | Building any mock — the HTML/CSS recipe, token mapping, surface-constraint table, fidelity specs, and `prefers-reduced-motion` rule |
Claude builds interfaces that look like Claude built them. Cyan-on-dark hero sections. Cards in a grid, all the same size. Glassmorphism on everything. Bounce easing. Inter at 16px. You've seen it. You've shipped it. Your users noticed.
Decomposes underspecified design requests by classifying gaps (missing brand context, visual ambiguity, false premises, scope…
Applies data visualization and information design principles to encode data truthfully and efficiently. Use when the question is…
Adjudicate whether users can operate an interface, and run heuristic evaluation. Use when the question is operability rather than…