agentleFS
RealTaskRealTask/MAX-DESIGN/SKILL.md

max-design

Skill4 starsChanged 23 days agoMITView on GitHub

A complete UI/UX engineering skill that combines frontend design, information architecture, UX writing, forms, design systems, data visualization, motion, accessibility, localization resilience, design review, and UI verification with bundled scripts and searchable design knowledge.

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use MAX-DESIGN for interface creation, product redesign, frontend implementation, UX improvements, design-system work, forms, dashboards, …

    line 12
  2. Distinctive:

    Build real behavior: Do not fake backend success, customer metrics, testimonials, saved states, or other unsupported product claims.

    line 20
  3. Distinctive:

    Make accessibility and resilience part of implementation: Consider semantic HTML, keyboard behavior, visible focus, contrast, reduced …

    line 21
  4. Distinctive:

    Read repository instructions, routes, representative screens, shared components, styles and tokens, assets, package scripts, tests, and …

    line 32
  5. Very specific:

    Run from the installed MAX-DESIGN directory:

    line 38
  6. Very specific:

    Read only the playbooks relevant to the task, then combine their guidance into one coherent implementation.

    line 79

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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

MAX-DESIGN

One integrated, production-minded UI/UX workflow for AI coding agents. This root skill contains the shared workflow, every specialist playbook, decision documents, examples, validation scripts, and the local searchable design-knowledge catalog.

When to use

Use MAX-DESIGN for interface creation, product redesign, frontend implementation, UX improvements, design-system work, forms, dashboards, accessibility fixes, responsive behavior, localization, motion, design critique, or quality verification. Use the full workflow for substantial or cross-cutting tasks and only the relevant stages for narrow tasks.

Core principles

  • Understand before styling. Inspect the existing product, routes, components, tokens, assets, tests, dependencies, and project instructions before changing code.
  • Make intentional design decisions. Choose layout, hierarchy, density, typography, color roles, and signature details for the user's task instead of defaulting to generic templates.
  • Preserve what works. Extend the existing stack and design system unless evidence supports a migration.
  • Design the whole state model. Include relevant loading, pending, empty, no-results, validation, error, retry, disabled, focus, success, and stale-data states.
  • Build real behavior. Do not fake backend success, customer metrics, testimonials, saved states, or other unsupported product claims.
  • Make accessibility and resilience part of implementation. Consider semantic HTML, keyboard behavior, visible focus, contrast, reduced motion, zoom, touch input, long content, localization, RTL where relevant, and recovery from failed requests.
  • Verify and report honestly. Separate code inspection, automated checks, browser testing, accessibility testing, performance measurement, and user research. Never claim a check that was not performed.

Workflow

1. Discover the user job

Identify the audience, primary task, primary action, important content, constraints, supported devices and input methods, risks, and observable acceptance criteria. Ask a question only when the answer could materially change the design; otherwise state a reasonable assumption and proceed.

2. Inspect the existing project

Read repository instructions, routes, representative screens, shared components, styles and tokens, assets, package scripts, tests, and existing documentation. Determine whether to preserve, extend, propose, or migrate the design system. Avoid unnecessary dependencies and broad rewrites.

3. Search the bundled design knowledge

Use the local catalog when a product, visual direction, page pattern, color, typography, motion, UX/accessibility, chart, implementation-stack, or density decision would benefit from a structured starting point. It is guidance, not a preset or an instruction to copy a pattern blindly. The knowledge is organized around original MAX-DESIGN records rather than imported third-party catalog files.

Run from the installed MAX-DESIGN directory:

node scripts/search-design-knowledge.mjs "dashboard charts"
node scripts/search-design-knowledge.mjs "AI assistant output states" --domain product
node scripts/search-design-knowledge.mjs "dark dashboard palette" --domain color
node scripts/search-design-knowledge.mjs "form errors" --domain ux
node scripts/search-design-knowledge.mjs "reduced motion" --domain motion
node scripts/search-design-knowledge.mjs "chart accessibility" --domain chart
node scripts/search-design-knowledge.mjs "responsive charts" --domain stack
node scripts/search-design-knowledge.mjs "responsive charts" --require-all --limit 5
node scripts/search-design-knowledge.mjs "modal focus lifecycle" --domain ux
node scripts/search-design-knowledge.mjs "offline retry" --domain ux
node scripts/search-design-knowledge.mjs "subscription invoice" --domain page
node scripts/search-design-knowledge.mjs "Core Web Vitals" --domain stack

Available domains: all, direction, page, interaction, density, product, style, color, typography, motion, ux, chart, and stack. Use --json for machine-readable results. The search loader merges knowledge/design-patterns.json and knowledge/advanced-patterns.json; the search implementation is at scripts/search-design-knowledge.mjs. The supplementary records include advanced product workflows, AI streaming, account and billing, privacy and permissions, resilient network states, ARIA widget behavior, WCAG 2.2 checks, chart data integrity, and evidence-based performance review. Research provenance and source limitations are documented in docs/knowledge-sources.md. Search remains local and requires no network connection or third-party runtime dependency.

4. Select a specific direction

Read docs/decision-framework.md before substantial design work. Define the product personality, content hierarchy, focal point, type roles, semantic color, density, responsive behavior, one or two product-relevant signature details, motion rationale, and trade-offs. Do not default to gradients, glass, giant radii, repeated card grids, oversized headlines, or animated backgrounds without product-specific reasons.

5. Load the relevant specialist playbooks

MAX-DESIGN is the single installable skill. Its specialist workflows are bundled as reference playbooks in skills/; read the relevant files when needed. They are not intended to be installed as separate peer skills by the standard repository installation.

TaskBundled playbook
Distinctive frontend and responsive implementationskills/frontend-design/SKILL.md
Navigation, information hierarchy, search, and task flowsskills/information-architecture/SKILL.md
Labels, onboarding, helper text, errors, and UX copyskills/ux-writing-content/SKILL.md
Forms, validation, submission, and recoveryskills/forms-ux/SKILL.md
Tokens, components, themes, and safe migrationsskills/design-system/SKILL.md
Charts, metrics, tabular data, and analytical displaysskills/data-visualization/SKILL.md
Evidence-based interface critiqueskills/design-review/SKILL.md
Transitions, micro-interactions, and motionskills/interaction-motion/SKILL.md
Semantics, keyboard, focus, contrast, and accessibilityskills/accessibility-audit/SKILL.md
Localization, long text, RTL, and resilient error pathsskills/resilient-localization/SKILL.md
Interaction, responsive, accessibility, and visual verificationskills/ui-testing/SKILL.md

Read only the playbooks relevant to the task, then combine their guidance into one coherent implementation. Shared principles and decision rules live in docs/design-principles.md, docs/quality-matrix.md, and docs/preflight-checklist.md.

6. Plan and implement

Structure content around the user task. Use semantic, maintainable components and the project's established conventions. Every visible control must perform a meaningful action, communicate its true state, or be explicitly disabled. Use realistic content and data. Preserve entered information after recoverable failures; provide clear feedback, retry paths, and confirmations proportionate to risk.

7. Review in impact order

Fix task-blocking behavior, incorrect information, confusing hierarchy, inaccessible interactions, broken responsive layouts, missing recovery states, and inconsistent tokens before decorative polish. Prioritize findings using docs/quality-matrix.md; separate observed issues from subjective preferences or untested assumptions.

8. Verify proportionately

Run the project's existing build, tests, lint, type checks, and other relevant scripts. Inspect narrow, medium, and wide layouts and exercise the key paths when browser tools are available. Use automated accessibility checks alongside manual keyboard review where practical. Check realistic long labels, empty states, errors, missing data, slow requests, and reduced-motion preferences when relevant.

The bundled repository has a dependency-free validator (Node.js 20+):

npm test

The design-knowledge search is also available through:

npm run search:design -- "responsive charts"

The validator checks the specialist playbooks, shared references, example tokens, catalog structure, and search smoke tests. Run it from the MAX-DESIGN directory. Do not run repository maintenance scripts from a target project unless you intend to validate the installed toolkit itself.

9. Hand off with evidence

For substantial work, summarize what changed, why major design decisions were made, what commands and viewports were checked, what remains unverified, and any meaningful trade-offs or risks. Never overstate accessibility, browser compatibility, performance, or conformance results.

Quality gates

  • User goals, information hierarchy, and primary actions are clear.
  • The design direction is distinctive for the product, not decorative by default.
  • Existing project conventions and working behavior are preserved.
  • UX copy, forms, charts, navigation, and data states are accurate and recoverable where relevant.
  • Responsive behavior, keyboard operation, focus, semantics, contrast, zoom, and reduced motion are considered.
  • Long text, localization, missing data, failed requests, and retry behavior are handled when applicable.
  • Performance and dependency costs are proportionate to the value delivered.
  • Verification claims match actual evidence and limitations are explicit.

Bundled content map

  • docs/ — decision framework, design principles, quality matrix, and preflight checks.
  • examples/ — design brief template and semantic design-token example.
  • knowledge/design-patterns.json — local searchable catalog of design directions, page patterns, interaction patterns, and density profiles.
  • scripts/search-design-knowledge.mjs — offline catalog-search CLI.
  • scripts/validate-skills.mjs — toolkit structure, metadata, token, catalog, and search validation.
  • skills/ — specialist playbooks included for on-demand use within this one MAX-DESIGN skill.

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.