notque/vexjoy-agent / skills/frontend/jev-design
jev-design
VerifiedDesign React interfaces with the full shadcn/ui catalog. The agent writes a brief, Jev picks components and a style recipe from bounded catalogs, then the agent implements, renders, and reviews against concrete checks.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 7.4 KB |
| references/brief-writing.md | markdown | 3.3 KB |
| references/catalog-v1.json | config | 2.3 KB |
| references/catalog-v2.json | config | 48.9 KB |
| references/decision-card.md | markdown | 2.5 KB |
| references/jev-images.md | markdown | 3.9 KB |
| references/visual-review.md | markdown | 3.3 KB |
| scripts/plan.py | script | 28.2 KB |
| tests/test_plan.py | script | 17.3 KB |
references/visual-review.md
View on GitHub# Visual review
Check the rendered screen, not the code. This checklist applies [UI design judgment](../../../shared-patterns/ui-design-judgment.md) section 10 to shadcn screens; run that section's page check script at each width too. Every item is pass or fail with a location. Default viewport widths are 375, 768, and 1280 px, each in light and dark mode.
## Tools
| Harness has | Do |
|---|---|
| Playwright MCP | `browser_navigate`, `browser_resize`, `browser_take_screenshot`, `browser_snapshot` (accessibility tree), `browser_press_key` (Tab order) |
| chrome-devtools MCP | `navigate_page`, `resize_page`, `take_screenshot`, `take_snapshot`, `emulate` (dark mode, reduced motion), `lighthouse_audit` (accessibility score) |
| neither | Say so. Give the user this checklist and the dev-server URL (bound to `127.0.0.1`). Do not claim visual quality. |
Toggle dark mode by adding the `dark` class to `<html>`, or by emulating `prefers-color-scheme: dark` when the app follows the system.
## Ask Jev about the screenshots
After the screenshots exist, send them to Jev as images (one Noul per trait) rather than describing them: see [jev-images.md](jev-images.md). Run the checklist below either way.
## Checklist
**Hierarchy**
- [ ] The screen has one `h1`, and it names the task or product.
- [ ] Within 5 seconds of looking at the screenshot, you can name the primary action. It is the only `default`-variant button in its region.
- [ ] Headings, labels, and numbers alone tell the story when the operator only scans.
**Spacing and alignment**
- [ ] Controls in the same row share one height (`layout_tokens.control`).
- [ ] Left edges align on a shared grid. No element is off by a few pixels.
- [ ] Gaps come from the plan's `gap` and `section` tokens, with no one-off values.
**Contrast and color**
- [ ] The plan's `contrast` rows all read `ok: true`. Custom colors you added reach 4.5:1 for text and 3:1 for icons, borders, and focus rings.
- [ ] One accent color. Red appears only for destructive or error states.
**Accessibility**
- [ ] Tab reaches every control in visual order, and focus is always visible.
- [ ] Every icon-only button has an accessible name (check the accessibility tree).
- [ ] Dialogs and sheets trap focus and return it to their trigger on close.
- [ ] Touch targets are at least 44 × 44 px at 375 px width.
- [ ] With reduced motion on, nothing slides or bounces.
**Responsive**
- [ ] No horizontal scroll at 375 px.
- [ ] Tables either scroll inside their own region or switch to a list at 375 px.
- [ ] The primary action is visible without scrolling at 375 px, or sticks to the bottom.
**Dark mode**
- [ ] No white panels or black text left behind. Charts and images stay legible.
- [ ] Borders are still visible on `input` fields.
**States**
- [ ] Loading shows `skeleton` or `spinner`. Empty shows `empty` with a next step. Errors show `alert` or field errors with a recovery action.
## Reporting a finding
Write each finding as: location, fact, width and mode, fix. For example: "Members list, row actions, 375 px light: the kebab button is 32 px tall; raise it to 44 px (`size-11`)."
Feed findings back as follows:
- Wrong component for a job: change that requirement's role or capabilities in `brief.json`, and rerun the planner.
- Wrong tone: change `style_traits`, and rerun.
- Anything else: fix the code.
Stop after three rounds, and list any open findings.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |