agentleFS
dembrandtdembrandt/dembrandt-skills/skills/dembrandt/SKILL.md

dembrandt

Skill194 starsChanged 4 months agoMITView on GitHub

Orchestrator that routes a UI/UX task through brand, tokens, layout, components, polish and accessibility. Use for multi-concern work such as a design review, UI build or interface audit.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Gate: All interactive states handled.

    line 131
  2. Distinctive:

    Run only the stages the task needs.

    line 173

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

dembrandt — UX Pipeline Orchestrator

Concept by @VictorGjn.

Routes multi-concern UI/UX tasks through six ordered stages. Each stage loads sub-skills on demand — only what the task actually needs.

When to use this skill vs a sub-skill directly:

  • Multi-concern task ("design review", "audit interface", "build UI") → use this orchestrator
  • Single-concern task ("check my colour palette", "review button states") → go directly to the sub-skill
  • Brand-to-token-to-spec pipeline with a URL or DESIGN.md → use generate-ui-from-brand instead
  • Reproducing an existing page rather than designing one → use clone-website instead
  • Reading the tokens off a live site and nothing more → use extract-design instead

Pipeline

Stage 1 — Brand Foundation

Establish the visual language before making any token decisions.

Sub-skills (load as needed):

  • brand-visual-language — shape language, icon style, typography tone
  • algorithmic-color-palette — derive states and brand-tinted greys from brand colours
  • color-mode-and-theme — light vs dark vs combined, when to offer a theme selector

Gate: Brand tone and colour system agreed before proceeding.


Stage 2 — Design Tokens & Scales

Pin the numeric system so all components share a common foundation.

Sub-skills (load as needed):

  • modular-scale-typography — ratio-based type scales, minimum sizes, context-aware usage
  • elevation-and-depth — shadow scale, border-radius, card and modal patterns
  • surface-separation — one means per surface: border, elevation or fill; opacity is for state
  • button-states — six states: rest, hover, active, focus, disabled, loading
  • component-family-consistency — buttons, inputs, pills: shared radius, colour, height
  • sizing-units — px vs rem vs relative: which values move when the user enlarges text
  • status-colors-and-errors — minimal semantic colours, error recovery, prevention

Gate: Tokens defined and consistent across component family.


Stage 3 — Layout & Structure

Apply layout decisions to the specific product context.

Sub-skills (load as needed):

  • layout-paradigms-and-consistency — choose the layout paradigm that fits the content; reuse the same page skeleton across screens (macro-scale consistency)
  • gestalt-ui-organisation — group related controls: proximity, similarity, common region
  • visual-emphasis-and-hierarchy — one CTA per view, colour and size as emphasis
  • information-architecture — naming, mental models, data UI, confirm dialogs
  • ui-context-and-scope — hierarchy, breadcrumbs, colour regions, scope communication
  • marketing-vs-product-system — site and product differ in density, share one token layer
  • responsive-paradigms — mobile/tablet/desktop: nav, sections, sticky behaviour
  • ui-density — match density to platform and user type
  • sticky-and-fixed-elements — headers, bottom toolbars, z-index tokens
  • scroll-areas — avoid inner scroll, one axis only, user-controlled

Gate: Layout is coherent across breakpoints and user contexts.


Stage 4 — Components & Interaction

Review component patterns and interactive states.

Sub-skills (load as needed):

  • real-world-metaphors — cards, carousels, drawers: when to use and how
  • tab-navigation — tab types, overflow, keyboard nav, ARIA, state persistence
  • modal-and-overlay-patterns — tooltip/popover/drawer/modal hierarchy, focus management, destructive confirm
  • form-design — helper text, placeholder, validation, submit state
  • data-display-and-selection — grid/list/table, large hit areas, mass actions
  • repeated-component-alignment — repeated components as slot models: equal size, pinned anchors, clamp + recover overflowing text
  • operational-expert-tool-ui — dense, workflow-driven UIs for trained daily B2B users
  • coordinated-data-views — keep a table and a visual view (map, diagram, chart) synchronized
  • domain-expert-configuration — expose solver/algorithm settings in domain language
  • authentic-product-representation — real content and real output over staged mockups and marketing chrome
  • app-shell — top bar, app launcher, tenant and environment cue, status bar: one shell across an estate
  • global-toolbar-controls — currency, language, region and unit selectors
  • notifications-and-recovery — toasts, banners, retry, undo — always a path forward

Gate: All interactive states handled; no dead ends.


Stage 5 — UX Polish

Apply UX principles and motion to sharpen perceived quality.

Sub-skills (load as needed):

  • nielsen-usability-heuristics — 10 usability principles with review checklists
  • user-flows-and-guided-paths — wizards, purchase flows, onboarding sequences
  • micro-interactions — animated icons, toggles, reveals, celebrations
  • loading-states-and-perceived-performance — spinners, skeleton screens, staggered entry
  • motion-and-storytelling — Disney principles and cinematic language in UI

Gate: Flow is legible end-to-end; perceived performance is acceptable.


Stage 6 — Accessibility & Technical Gate

Hard ship gate. Do not skip or defer.

Sub-skills (load as needed):

  • wcag-accessibility — WCAG 2.2 AA / EN 301 549: contrast, keyboard, ARIA
  • semantic-html-and-seo — HTML5, alt texts, Open Graph, progressive enhancement
  • performance-and-web-vitals — Lighthouse audit, LCP, CLS, INP, images, fonts, JS loading

Gate: Passes WCAG 2.2 AA. Required by EU Accessibility Act (EAA) for products launched after June 2025.


Output

Produce one structured review, not a stream of loose comments. Group findings by severity, most blocking first:

  1. Blockers — fails a hard gate: WCAG 2.2 AA, a broken flow, a dead end. Must fix before ship.
  2. Major — breaks consistency, hierarchy, or a core UX principle. Fix this iteration.
  3. Minor — polish, micro-interactions, perceived-performance refinements. Backlog.

For each finding give the stage it came from, the specific element, what is wrong, and the concrete fix.

Run only the stages the task needs. An existing product with settled brand and tokens starts at Stage 3; a pure accessibility pass runs Stage 6 alone. State which stages you ran and which you skipped, and why.


Relationship to generate-ui-from-brand

generate-ui-from-brand is a token-extraction pipeline: URL or DESIGN.md → tokens → UI spec. It overlaps stages 1–2 of this orchestrator. Use it when you have a brand source and need a concrete spec. Use this orchestrator when you are reviewing or building across the full stack of UX concerns without a specific brand-extraction starting point.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.