[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7AR5X7BKT9ABBBAM65BNN2N":3,"skill-same:01M7AR5X7BKT9ABBBAM65BNN2N":64},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":30,"commitSha":31,"skillDirectory":32,"forksCount":14,"repositoryHtmlUrl":33,"repositoryHomepageUrl":34,"ownerHtmlUrl":35,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":36,"install":37,"resources":49},"01M7AR5X7BKT9ABBBAM65BNN2N","brand-ui-audit",null,"Audit and refine UI built with brand-ui — visual quality, design consistency, accessibility, and token compliance. Use when the user asks to audit\u002Freview\u002Fpolish a UI, check contrast or accessibility, find design issues, make an interface feel more finished, verify it works across themes, or before shipping a screen. Runs a deterministic static pass (token violations, spacing, sizing, anti-patterns) plus a rendered cross-theme visual + WCAG-contrast pass over Storybook or a running app, then produces a phased, token-referenced plan. Read-only by default — reports findings; every fix points at a @brand token, never raw hex.","mreimitz\u002Felabs-components","mreimitz","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F27810221?v=4","01M7AR5BCPZJ6V2RNWHAZTS47W","Manuel Reimitz","01M7AR5BDN5NMH80W8GCXACXXD",1,"MIT","https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-audit\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: brand-ui-audit\ndescription: Audit and refine UI built with brand-ui — visual quality, design consistency, accessibility, and token compliance. Use when the user asks to audit\u002Freview\u002Fpolish a UI, check contrast or accessibility, find design issues, make an interface feel more finished, verify it works across themes, or before shipping a screen. Runs a deterministic static pass (token violations, spacing, sizing, anti-patterns) plus a rendered cross-theme visual + WCAG-contrast pass over Storybook or a running app, then produces a phased, token-referenced plan. Read-only by default — reports findings; every fix points at a @brand token, never raw hex.\nuser-invocable: true\nargument-hint: \"[audit|contrast|review|polish] [target]\"\nallowed-tools:\n  - Bash(npx @elabs-ai\u002Fcomponents-cli *)\n  - Bash(pnpm brand-ui *)\n  - Bash(npx brand-ui *)\n---\n\n# brand-ui-audit\n\nA design-quality reviewer for brand-ui interfaces. It catches what static checks\nand unit tests miss — weak hierarchy, cramped spacing, low contrast, token\nviolations, broken theming, anti-patterns — and reports them as a prioritized,\nimplementation-ready plan. **Read-only**: it reports and (with approval) files\nissues; the fix is done separately.\n\n> Run the CLI with the project's runner. In this monorepo: `pnpm brand-ui \u003Ccmd>`.\n> In a consuming project, add it with `pnpm add -D @elabs-ai\u002Fcomponents-cli` (a\n> public npm package — no registry setup, no token), then run\n> `pnpm exec brand-ui \u003Ccmd>`; `npx -y @elabs-ai\u002Fcomponents-cli \u003Ccmd>` works with no\n> install at all. **Inside Claude Code with the plugin installed, prefer the\n> `mcp__brand-ui__*` tools** (`mcp__brand-ui__info`, `mcp__brand-ui__audit`) — the\n> plugin launches the same CLI for you, so you skip the per-project\n> dev-dependency. Examples below say `brand-ui`.\n\n## Setup\n\n1. Run **`brand-ui info`** once (via `mcp__brand-ui__info`, or the runner from the\n   note above) to load the theme list and token set. Every\n   recommendation must resolve to a token from this set — never propose raw hex.\n2. Identify the target: a file\u002Fdir (static pass), a Storybook story id, or a\n   running app URL (rendered pass).\n3. **Read the active taste profile — don't ask for it.** The same `brand-ui info`\n   call carries a `taste` block: `register × density × motion × expressiveness`\n   (brand-ui ADR 0020). It resolves the shipped\n   defaults against the optional `brand-ui.config.json` `taste` key **nearest the\n   code being judged** (the audit target's own ancestors first, then the cwd, then\n   the repo root). **If it is absent** (an old CLI, or a project with no config), fall back to\n   the restrained default — `product \u002F comfortable \u002F system \u002F 0` — and **say so in\n   the report**, so nobody mistakes a default for a decision. Only ask the user\n   when the surface in focus plainly contradicts the resolved register (a landing\n   page in a `product` repo — judge by the surface in focus, not the whole repo)\n   and say which one you used. The register flips the defaults you audit against:\n   - **product** (the default for brand-ui) — app UI, dashboards, admin, tools. Bar is\n     _earned familiarity_: one type family, fixed rem scale, restrained color,\n     every component state present (default\u002Fhover\u002Ffocus\u002Factive\u002Fdisabled\u002Floading\u002F\n     empty\u002Ferror), skeletons over spinners, quick 150–250ms state-only motion, no\n     modal-as-first-thought. Most `@elabs-ai\u002Fcomponents-*` surfaces are here.\n   - **brand** — `@elabs-ai\u002Fcomponents-marketing` surfaces, landing pages, campaigns. Bar is\n     _distinctiveness_: committed color, required real imagery (not colored blocks),\n     ambitious first-load motion, a POV. \"Restraint without intent reads as\n     mediocre.\"\n\n   The deterministic pass honours the register for you — `brand-ui audit` softens\n   `over-round` \u002F `side-stripe` \u002F `bounce-easing` to advisory in the `brand`\n   register (override with `--register=product|brand` when you audit one surface\n   against a different bar than the project's). It never softens a banned rule\n   (raw color, `gradient-text`, tiny text) or content slop.\n\n## Modes (route on the first argument)\n\n| Mode                  | Does                                                                          |\n| --------------------- | ----------------------------------------------------------------------------- |\n| `audit [path]`        | Full pass: static lint **+** rendered cross-theme visual + contrast. Default. |\n| `contrast [target]`   | Just the WCAG pass (oklch-aware) across all themes.                           |\n| `review [story\\|url]` | Visual\u002FUX critique of a specific surface (hierarchy, spacing, states).        |\n| `polish [target]`     | Re-audit after fixes; confirm findings are resolved, no regressions.          |\n\n## Pass 1 — deterministic (no browser, no LLM)\n\nRun `brand-ui audit \u003Cpath> --json` (via `mcp__brand-ui__audit`, or the runner from\nthe note above). It flags, per file:line:\n\n- **raw-hex \u002F rgb-literal \u002F arbitrary-color** — colors outside `themes.css`\n  (the one allowed place). Fix: a semantic token (`bg-card`, `text-muted-foreground`, …).\n- **space-y-x** — `space-x\u002Fspace-y-*` → use `flex`\u002F`grid` + `gap-*`.\n- **wh-equal** — `w-N h-N` with equal values → `size-N`.\n- **outline-none** — focus ring possibly removed → keep `focus-visible:ring-2 ring-ring`.\n- **Visual anti-slop (WP-15)** — `pure-black` (use the `foreground` token),\n  `neon-glow` (tinted\u002Finset shadow), `gradient-text`, `side-stripe`, `over-round`,\n  `custom-cursor`, `viewport-h-screen` (`min-h-dvh`). Token-translated tells.\n- **Composition** — `ai\u002Fprefer-composer`: a direct `\u003CPromptInput>` render where\n  `\u003CComposer>` is the canonical chat input. **Advisory, always** — dropping to the\n  primitive for a bespoke shell (a mention roster, say) is a documented escape\n  hatch, so this never fails `--strict`. Quiet in a file that also renders\n  `\u003CComposer>`, in the file that defines it, in the `PromptInput` family's own\n  modules\u002Fstories, and in any test. A file argues its own case with\n  `\u002F\u002F brand-ui-audit-allow: ai\u002Fprefer-composer` — which works on advisory rules\n  only, never on a blocking one.\n- **Content anti-slop — the \"Jane Doe effect\" (WP-15)** — `slop-generic-name`\n  (\"John\u002FJane Doe\"), `slop-fake-number` (\"99.99%\"), `slop-brand-name`\n  (\"Acme\u002FNexus\"), `marketing-buzzword` (filler verbs). Advisory here (placeholders\n  can be intentional); the audit reports them, and a reviewer decides.\n\nThese are cheap, exact, and CI-friendly. The register-gated\u002Fperceptual tells\n(anti-card, 3-equal-cards, motion intensity) and rendered contrast stay in Pass 2.\n\n## Pass 2 — rendered (cross-theme visual + contrast)\n\nThis is the high-signal pass and needs a browser. Use the **agent-browser** skill\nto drive a running Storybook (default `http:\u002F\u002Flocalhost:6007`) or app:\n\n1. Enumerate the surfaces to check (Storybook `\u002Findex.json`, or the app's routes).\n   Pick a representative set: app shell, data table, chat, charts, flow, forms,\n   overlays (opened), states (empty\u002Floading\u002Ferror), plus foundation (button, badge, alert).\n2. For **each theme** (light, dark): navigate with a render\n   gate, screenshot, and read the pixels. Switch a Storybook story's theme via\n   `&globals=theme:\u003Clight|dark>`.\n   **Always wait for render before the screenshot** (poll for content \u002F fixed\n   delay) — a screenshot fired during the loader is a capture bug, not a finding.\n3. **Measure contrast on the real pixels**, oklch-aware. Computed colors here come\n   back as `oklch()`, so an `rgb()`-only parser silently passes everything —\n   convert oklch → sRGB → relative luminance, then ratio. Flag body text \u003C 4.5:1,\n   large text\u002FUI \u003C 3:1. (This is exactly how the brand green was caught at\n   3.61:1 white-on-green and 3.77:1 green-on-white.)\n4. Check focus rings (tab through), hover\u002Factive\u002Fdisabled, and that each theme\n   actually repaints (`document.documentElement.dataset.theme`).\n\nSee [reference\u002Fcontrast-audit.md](reference\u002Fcontrast-audit.md) for the oklch\ncontrast function and the capture-gate recipe, and\n[reference\u002Fanti-patterns.md](reference\u002Fanti-patterns.md) for the design smells.\n\n**Two ways to measure rendered contrast** (the reference covers the first; the\nsecond is the more robust upgrade):\n\n1. **Computed-color** (default here): read each text element's computed `color` +\n   effective background, convert oklch→sRGB→luminance, ratio. Accurate for the\n   solid-token backgrounds brand-ui normally uses; blind to gradients\u002Fimages\n   behind text.\n2. **Screenshot-diff** (upgrade, for gradient\u002Fimage backgrounds): screenshot the\n   element, hide the glyph text (`color: transparent`), screenshot again,\n   pixel-diff to isolate glyph pixels, and take the **p10 percentile** contrast\n   against the actually-rendered background. This is how a mature detector\n   (impeccable's `screenshot-contrast.mjs`) avoids missing low contrast over a\n   photo or gradient. Reach for it when text sits on imagery.\n\n## Evaluate (scored)\n\nRun **two independent passes then synthesize** — form the design\u002Fheuristic read\n(Pass A) before the detector numbers (Pass B) enter judgment, so the deterministic\noutput doesn't anchor the eye. Then score with the rubric in\n[reference\u002Fux-evaluation.md](reference\u002Fux-evaluation.md):\n\n- **\u002F24 scorecard** (accessibility · states & resilience · theming & tokens ·\n  consistency & hierarchy · visual anti-patterns · **taste & anti-slop**), optional\n  0–100 composite + a one-line \"does this look AI-generated?\" verdict that now\n  weighs **content** slop (the \"Jane Doe effect\"), not just visual tells. The taste\n  axis is **register-gated** — judge against the **resolved** taste profile from\n  Setup step 3 (`brand-ui info` → `taste`) and state it in the report; product is\n  restrained, brand is expressive.\n- **9-state inventory** per component (default\u002Fhover\u002Ffocus\u002Factive\u002Fdisabled\u002Floading\u002F\n  empty\u002Ferror\u002Fpartial) — flag missing or ad-hoc states; check brand-ui's\n  `EmptyState`\u002F`Skeleton`\u002F`ErrorState`\u002F`LoadingState` are used, not bespoke markup.\n- **WCAG 2.2 for designers** (POUR + the 2.2 additions: 24px targets, focus not\n  obscured, redundant entry, dragging alternative).\n- **Copy & microcopy** (error = what+how, empty = what+how+why, verb+object\n  buttons, destructive friction hierarchy), the **content anti-slop** check (the\n  \"Jane Doe effect\"), and an **ethics \u002F dark-pattern** scan.\n- **Taste pre-flight** (ux-evaluation.md) — the final, register-gated QA gate\n  (tokens-not-literals, real content, every state, motion honored-not-mandated,\n  cross-theme contrast, hierarchy\u002Flayout) before a surface is \"done\".\n\nApply the **reduction filter** — if an element can be removed without losing\nmeaning, recommend removing it; if a user would need to be told it exists, rework it.\n\n## Output — phased plan, then approval\n\nGroup findings by severity and present as a plan; **do not edit components**:\n\n- **P0** — broken \u002F illegible \u002F inaccessible (e.g. AA contrast failures).\n- **P1** — clearly hurts quality.\n- **P2** — polish.\n\nEach finding: surface + theme, what's wrong, why it matters, and a **concrete,\ntoken-referenced fix** (name the token or rule). Write the report to a dated\nfile (e.g. `reports\u002Fvisual-ux-\u003Cdate>.md`) with screenshot links.\n\nThen **report the findings to the user** — for each: `file:line` (or\nsurface + theme), severity, and the concrete, token-referenced fix. This skill\n**reports, it does not fix**: the user decides what to change. Don't edit\ncomponents as part of the audit.\n\n## Token discipline (what makes this on-brand, not generic)\n\nNever recommend a raw color, a hardcoded radius, or an arbitrary value. Every fix\nis a `@brand` token or a documented rule. If a needed visual concept has no token,\nsay so and route to `brand-ui-theme` to add it — don't invent a literal.\n","# brand-ui-audit\n\nA design-quality reviewer for brand-ui interfaces. It catches what static checks\nand unit tests miss — weak hierarchy, cramped spacing, low contrast, token\nviolations, broken theming, anti-patterns — and reports them as a prioritized,\nimplementation-ready plan. **Read-only**: it reports and (with approval) files\nissues; the fix is done separately.\n\n> Run the CLI with the project's runner. In this monorepo: `pnpm brand-ui \u003Ccmd>`.\n> In a consuming project, add it with `pnpm add -D @elabs-ai\u002Fcomponents-cli` (a\n> public npm package — no registry setup, no token), then run\n> `pnpm exec brand-ui \u003Ccmd>`; `npx -y @elabs-ai\u002Fcomponents-cli \u003Ccmd>` works with no\n> install at all. **Inside Claude Code with the plugin installed, prefer the\n> `mcp__brand-ui__*` tools** (`mcp__brand-ui__info`, `mcp__brand-ui__audit`) — the\n> plugin launches the same CLI for you, so you skip the per-project\n> dev-dependency. Examples below say `brand-ui`.\n\n## Setup\n\n1. Run **`brand-ui info`** once (via `mcp__brand-ui__info`, or the runner from the\n   note above) to load the theme list and token set. Every\n   recommendation must resolve to a token from this set — never propose raw hex.\n2. Identify the target: a file\u002Fdir (static pass), a Storybook story id, or a\n   running app URL (rendered pass).\n3. **Read the active taste profile — don't ask for it.** The same `brand-ui info`\n   call carries a `taste` block: `register × density × motion × expressiveness`\n   (brand-ui ADR 0020). It resolves the shipped\n   defaults against the optional `brand-ui.config.json` `taste` key **nearest the\n   code being judged** (the audit target's own ancestors first, then the cwd, then\n   the repo root). **If it is absent** (an old CLI, or a project with no config), fall back to\n   the restrained default — `product \u002F comfortable \u002F system \u002F 0` — and **say so in\n   the report**, so nobody mistakes a default for a decision. Only ask the user\n   when the surface in focus plainly contradicts the resolved register (a landing\n   page in a `product` repo — judge by the surface in focus, not the whole repo)\n   and say which one you used. The register flips the defaults you audit against:\n   - **product** (the default for brand-ui) — app UI, dashboards, admin, tools. Bar is\n     _earned familiarity_: one type family, fixed rem scale, restrained color,\n     every component state present (default\u002Fhover\u002Ffocus\u002Factive\u002Fdisabled\u002Floading\u002F\n     empty\u002Ferror), skeletons over spinners, quick 150–250ms state-only motion, no\n     modal-as-first-thought. Most `@elabs-ai\u002Fcomponents-*` surfaces are here.\n   - **brand** — `@elabs-ai\u002Fcomponents-marketing` surfaces, landing pages, campaigns. Bar is\n     _distinctiveness_: committed color, required real imagery (not colored blocks),\n     ambitious first-load motion, a POV. \"Restraint without intent reads as\n     mediocre.\"\n\n   The deterministic pass honours the register for you — `brand-ui audit` softens\n   `over-round` \u002F `side-stripe` \u002F `bounce-easing` to advisory in the `brand`\n   register (override with `--register=product|brand` when you audit one surface\n   against a different bar than the project's). It never softens a banned rule\n   (raw color, `gradient-text`, tiny text) or content slop.\n\n## Modes (route on the first argument)\n\n| Mode                  | Does                                                                          |\n| --------------------- | ----------------------------------------------------------------------------- |\n| `audit [path]`        | Full pass: static lint **+** rendered cross-theme visual + contrast. Default. |\n| `contrast [target]`   | Just the WCAG pass (oklch-aware) across all themes.                           |\n| `review [story\\|url]` | Visual\u002FUX critique of a specific surface (hierarchy, spacing, states).        |\n| `polish [target]`     | Re-audit after fixes; confirm findings are resolved, no regressions.          |\n\n## Pass 1 — deterministic (no browser, no LLM)\n\nRun `brand-ui audit \u003Cpath> --json` (via `mcp__brand-ui__audit`, or the runner from\nthe note above). It flags, per file:line:\n\n- **raw-hex \u002F rgb-literal \u002F arbitrary-color** — colors outside `themes.css`\n  (the one allowed place). Fix: a semantic token (`bg-card`, `text-muted-foreground`, …).\n- **space-y-x** — `space-x\u002Fspace-y-*` → use `flex`\u002F`grid` + `gap-*`.\n- **wh-equal** — `w-N h-N` with equal values → `size-N`.\n- **outline-none** — focus ring possibly removed → keep `focus-visible:ring-2 ring-ring`.\n- **Visual anti-slop (WP-15)** — `pure-black` (use the `foreground` token),\n  `neon-glow` (tinted\u002Finset shadow), `gradient-text`, `side-stripe`, `over-round`,\n  `custom-cursor`, `viewport-h-screen` (`min-h-dvh`). Token-translated tells.\n- **Composition** — `ai\u002Fprefer-composer`: a direct `\u003CPromptInput>` render where\n  `\u003CComposer>` is the canonical chat input. **Advisory, always** — dropping to the\n  primitive for a bespoke shell (a mention roster, say) is a documented escape\n  hatch, so this never fails `--strict`. Quiet in a file that also renders\n  `\u003CComposer>`, in the file that defines it, in the `PromptInput` family's own\n  modules\u002Fstories, and in any test. A file argues its own case with\n  `\u002F\u002F brand-ui-audit-allow: ai\u002Fprefer-composer` — which works on advisory rules\n  only, never on a blocking one.\n- **Content anti-slop — the \"Jane Doe effect\" (WP-15)** — `slop-generic-name`\n  (\"John\u002FJane Doe\"), `slop-fake-number` (\"99.99%\"), `slop-brand-name`\n  (\"Acme\u002FNexus\"), `marketing-buzzword` (filler verbs). Advisory here (placeholders\n  can be intentional); the audit reports them, and a reviewer decides.\n\nThese are cheap, exact, and CI-friendly. The register-gated\u002Fperceptual tells\n(anti-card, 3-equal-cards, motion intensity) and rendered contrast stay in Pass 2.\n\n## Pass 2 — rendered (cross-theme visual + contrast)\n\nThis is the high-signal pass and needs a browser. Use the **agent-browser** skill\nto drive a running Storybook (default `http:\u002F\u002Flocalhost:6007`) or app:\n\n1. Enumerate the surfaces to check (Storybook `\u002Findex.json`, or the app's routes).\n   Pick a representative set: app shell, data table, chat, charts, flow, forms,\n   overlays (opened), states (empty\u002Floading\u002Ferror), plus foundation (button, badge, alert).\n2. For **each theme** (light, dark): navigate with a render\n   gate, screenshot, and read the pixels. Switch a Storybook story's theme via\n   `&globals=theme:\u003Clight|dark>`.\n   **Always wait for render before the screenshot** (poll for content \u002F fixed\n   delay) — a screenshot fired during the loader is a capture bug, not a finding.\n3. **Measure contrast on the real pixels**, oklch-aware. Computed colors here come\n   back as `oklch()`, so an `rgb()`-only parser silently passes everything —\n   convert oklch → sRGB → relative luminance, then ratio. Flag body text \u003C 4.5:1,\n   large text\u002FUI \u003C 3:1. (This is exactly how the brand green was caught at\n   3.61:1 white-on-green and 3.77:1 green-on-white.)\n4. Check focus rings (tab through), hover\u002Factive\u002Fdisabled, and that each theme\n   actually repaints (`document.documentElement.dataset.theme`).\n\nSee [reference\u002Fcontrast-audit.md](reference\u002Fcontrast-audit.md) for the oklch\ncontrast function and the capture-gate recipe, and\n[reference\u002Fanti-patterns.md](reference\u002Fanti-patterns.md) for the design smells.\n\n**Two ways to measure rendered contrast** (the reference covers the first; the\nsecond is the more robust upgrade):\n\n1. **Computed-color** (default here): read each text element's computed `color` +\n   effective background, convert oklch→sRGB→luminance, ratio. Accurate for the\n   solid-token backgrounds brand-ui normally uses; blind to gradients\u002Fimages\n   behind text.\n2. **Screenshot-diff** (upgrade, for gradient\u002Fimage backgrounds): screenshot the\n   element, hide the glyph text (`color: transparent`), screenshot again,\n   pixel-diff to isolate glyph pixels, and take the **p10 percentile** contrast\n   against the actually-rendered background. This is how a mature detector\n   (impeccable's `screenshot-contrast.mjs`) avoids missing low contrast over a\n   photo or gradient. Reach for it when text sits on imagery.\n\n## Evaluate (scored)\n\nRun **two independent passes then synthesize** — form the design\u002Fheuristic read\n(Pass A) before the detector numbers (Pass B) enter judgment, so the deterministic\noutput doesn't anchor the eye. Then score with the rubric in\n[reference\u002Fux-evaluation.md](reference\u002Fux-evaluation.md):\n\n- **\u002F24 scorecard** (accessibility · states & resilience · theming & tokens ·\n  consistency & hierarchy · visual anti-patterns · **taste & anti-slop**), optional\n  0–100 composite + a one-line \"does this look AI-generated?\" verdict that now\n  weighs **content** slop (the \"Jane Doe effect\"), not just visual tells. The taste\n  axis is **register-gated** — judge against the **resolved** taste profile from\n  Setup step 3 (`brand-ui info` → `taste`) and state it in the report; product is\n  restrained, brand is expressive.\n- **9-state inventory** per component (default\u002Fhover\u002Ffocus\u002Factive\u002Fdisabled\u002Floading\u002F\n  empty\u002Ferror\u002Fpartial) — flag missing or ad-hoc states; check brand-ui's\n  `EmptyState`\u002F`Skeleton`\u002F`ErrorState`\u002F`LoadingState` are used, not bespoke markup.\n- **WCAG 2.2 for designers** (POUR + the 2.2 additions: 24px targets, focus not\n  obscured, redundant entry, dragging alternative).\n- **Copy & microcopy** (error = what+how, empty = what+how+why, verb+object\n  buttons, destructive friction hierarchy), the **content anti-slop** check (the\n  \"Jane Doe effect\"), and an **ethics \u002F dark-pattern** scan.\n- **Taste pre-flight** (ux-evaluation.md) — the final, register-gated QA gate\n  (tokens-not-literals, real content, every state, motion honored-not-mandated,\n  cross-theme contrast, hierarchy\u002Flayout) before a surface is \"done\".\n\nApply the **reduction filter** — if an element can be removed without losing\nmeaning, recommend removing it; if a user would need to be told it exists, rework it.\n\n## Output — phased plan, then approval\n\nGroup findings by severity and present as a plan; **do not edit components**:\n\n- **P0** — broken \u002F illegible \u002F inaccessible (e.g. AA contrast failures).\n- **P1** — clearly hurts quality.\n- **P2** — polish.\n\nEach finding: surface + theme, what's wrong, why it matters, and a **concrete,\ntoken-referenced fix** (name the token or rule). Write the report to a dated\nfile (e.g. `reports\u002Fvisual-ux-\u003Cdate>.md`) with screenshot links.\n\nThen **report the findings to the user** — for each: `file:line` (or\nsurface + theme), severity, and the concrete, token-referenced fix. This skill\n**reports, it does not fix**: the user decides what to change. Don't edit\ncomponents as part of the audit.\n\n## Token discipline (what makes this on-brand, not generic)\n\nNever recommend a raw color, a hardcoded radius, or an arbitrary value. Every fix\nis a `@brand` token or a documented rule. If a needed visual concept has no token,\nsay so and route to `brand-ui-theme` to add it — don't invent a literal.\n",{"name":5,"description":7,"allowed-tools":24,"argument-hint":28,"user-invocable":29},[25,26,27],"Bash(npx @elabs-ai\u002Fcomponents-cli *)","Bash(pnpm brand-ui *)","Bash(npx brand-ui *)","[audit|contrast|review|polish] [target]",true,"main","f327bf0c3cd1eb26e3d5de071c864afa989b6c72","skills\u002Fbrand-ui-audit","https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components","https:\u002F\u002Felabs-ai.com\u002F","https:\u002F\u002Fgithub.com\u002Fmreimitz","2026-10-07 20:39:57",{"source":8,"skillName":5,"commands":38,"prompt":48,"zipAvailable":29},[39,42,45],{"tool":40,"command":41},"npx","npx skills add mreimitz\u002Felabs-components --skill brand-ui-audit",{"tool":43,"command":44},"bunx","bunx skills add mreimitz\u002Felabs-components --skill brand-ui-audit",{"tool":46,"command":47},"pnpm","pnpm dlx skills add mreimitz\u002Felabs-components --skill brand-ui-audit","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-audit\u002FSKILL.md",[50,56,60],{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"reference\u002Fanti-patterns.md","REFERENCE","参考文档",8730,"https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-audit\u002Freference\u002Fanti-patterns.md",{"path":57,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":58,"sourceHtmlUrl":59},"reference\u002Fcontrast-audit.md",4800,"https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-audit\u002Freference\u002Fcontrast-audit.md",{"path":61,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":62,"sourceHtmlUrl":63},"reference\u002Fux-evaluation.md",11424,"https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-audit\u002Freference\u002Fux-evaluation.md",[65,72,79,86,93,100],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M7AR8KP3VE2Y0QRA5D1RQNPZ","brand-ui","Build UI with the brand-ui component system (@elabs-ai\u002Fcomponents-* packages — ui, data, ai, flow, maps, charts, marketing, editor, viewer, terminal, tokens, icons). Use when working in a project that depends on @elabs-ai\u002Fcomponents-ui or any @elabs-ai\u002Fcomponents-* package, when adding\u002Fcomposing components, building dashboards, data tables, AI\u002Fchat surfaces, React Flow canvases, MapLibre maps, code editors, file\u002Fdocument viewers, app shells, forms, or marketing sections, when theming with the token system, or when the user mentions brand-ui, @brand, light\u002Fdark themes, or \"our design system\". Provides live project context, the real component API, composition patterns, and the rules that keep components token-driven, accessible, and theme-safe.","https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":19},"01M7AR8DHBD6W72TY76VJ6Z6FT","brand-ui-update-theme","Use when an existing brand-ui theme family should be improved from new material — updated brand guidelines, a new logo, product screenshots, a live site, feedback like \"the green is off\" or \"match their new design system\" — e.g. \"update the qlik theme with these links\", \"refresh our theme from the new brand book\", \"\u002Fupdate-theme\". For a brand with no theme yet, use brand-ui-create-theme.","https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-update-theme\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":19},"01M7AR86YT6GTFP4X1KEQ3TKG2","brand-ui-theme","Create or retune a brand-ui theme (token work). Use when adding a new theme, re-branding the system to a customer's palette, adjusting global radius\u002Fspacing\u002Fsurfaces, or fixing token-level contrast across themes. Use when the user says \"new theme\", \"brand it for \u003Ccustomer>\", \"make corners less rounded\", \"dark theme\", \"adjust the tokens\", or \"our colors\". Every change is a token in a theme stylesheet — never a hardcoded value in a component.","https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-theme\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":19},"01M7AR7X95MZNM7YTSD56F38HQ","brand-ui-start","The front door for building UI with brand-ui — a router\u002Fconcierge that asks what you want to do and hands you to the right flow. Use when someone is starting out and unsure where to begin (\"help me with brand-ui\", \"where do I start\", \"I want to build something\", \"what can brand-ui do\", \"get me going\"), or types \u002Fbrand-ui-start. Routes to three flows — build a NEW app (brand-ui-new-app), improve\u002Fmigrate an EXISTING app (brand-ui-migrate), or just help me USE brand-ui in an app that already has it (brand-ui). If the intent is already obvious (the person said \"build me a dashboard\" or \"audit this screen\"), skip the question and go straight to that flow.","https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-start\u002FSKILL.md",[],[],{"skillNo":94,"name":95,"nameZh":6,"description":96,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":19},"01M7AR7MAA8KRQ4HDKPYGXDGG6","brand-ui-registry","Curate the brand-ui shadcn-compatible registry (maintainer workflow). Use when deciding whether something belongs as an imported @elabs-ai\u002Fcomponents-* primitive vs a copy-own registry block\u002Ftemplate, when adding or editing a registry item, or when validating\u002Fbuilding the registry for distribution. Use when the user says \"add a registry item\", \"make this a block\", \"publish to the registry\", or \"package vs registry\". For consuming registry items in an app, use the `brand-ui` skill (npx shadcn add).","https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-registry\u002FSKILL.md",[],[],{"skillNo":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":19},"01M7AR79ZG2N23Z7GAXGVAZKA7","brand-ui-new-app","Define-to-build — scaffold a new brand-ui app from a plain-language description (VP-02). Use when the user wants to START an app, page, or demo (\"build me a sales dashboard\", \"I need an admin console for X\", \"new app\", \"scaffold a chat assistant\", \"create a landing page for a pitch\") rather than add to an existing one. Runs a staged interview (quick 3-question mode or full 7-stage spec mode), writes an app-spec.md, then scaffolds from the matching template + playbook with every wiring point annotated and a starter CLAUDE.md so later agent sessions stay on-brand. For adding components to an existing app use `brand-ui`; for authoring library components use `brand-ui-component`.","https:\u002F\u002Fgithub.com\u002Fmreimitz\u002Felabs-components\u002Fblob\u002Fmain\u002Fskills\u002Fbrand-ui-new-app\u002FSKILL.md",[],[]]