ui-ux-master 1.2.0 → 1.5.0

package/SKILL.md

@@ -1,7 +1,7 @@
 ---
 name: ui-ux-master
 description: "Use when an agent must design, audit, improve, prototype, specify, or hand off any UI/UX work end-to-end. Provides one-stop UX strategy, research, IA, flows, wireframes, visual design, design systems, accessibility, usability testing, conversion, and engineering handoff workflows."
-version: 1.2.0
+version: 1.5.0
 author: Rupak Biswas
 license: MIT
 platforms: [linux, macos, windows]
@@ -23,32 +23,38 @@ The job is not only to make screens look good. The job is to solve the user's go
 
 This skill is opt-in. In Codex, Claude, Windsurf, Antigravity, Gemini, Cursor, and other agents, activate it only when the user includes `/ui-ux-master` in the prompt or invokes the installed native slash command. If the user does not mention `/ui-ux-master`, do not apply this skill automatically.
 
-When activated, the user can write naturally, for example: `/ui-ux-master audit this checkout flow` or `/ui-ux-master design a premium SaaS landing page`. Treat the rest of the prompt as the task and follow this SKILL.md plus the referenced files.
-
-## AI Discovery and MCP
-
-This package exposes the workflow through human-readable instructions and machine-readable discovery assets:
-
-- `llms.txt` gives AI agents a short map of the package.
-- `ai-discovery/ui-ux-master.manifest.json` describes activation, entrypoints, tools, resources, prompts, and supported agents.
-- `system-prompts/` contains full, compact, and MCP-focused system prompt add-ons.
-- `bin/ui-ux-master-mcp.mjs` runs a local read-only MCP server that exposes UI/UX Master tools, resources, and prompts.
-
+When activated, the user can write naturally, for example: `/ui-ux-master audit this checkout flow` or `/ui-ux-master design a premium SaaS landing page`. Treat the rest of the prompt as the task and follow this SKILL.md plus the referenced files.
+
+## AI Discovery and MCP
+
+This package exposes the workflow through human-readable instructions and machine-readable discovery assets:
+
+- `llms.txt` gives AI agents a short map of the package.
+- `ai-discovery/ui-ux-master.manifest.json` describes activation, entrypoints, tools, resources, prompts, and supported agents.
+- `system-prompts/` contains full, compact, and MCP-focused system prompt add-ons.
+- `bin/ui-ux-master-mcp.mjs` runs a local read-only MCP server that exposes UI/UX Master tools, resources, and prompts.
+
 Use MCP and system prompt add-ons only as discovery/activation layers. The workflow remains opt-in and should still require `/ui-ux-master` unless the user explicitly asks to use UI/UX Master.
 
 ## Core Operating Rule
 
 Do not jump directly to visuals. Work in this order unless the user explicitly asks for a narrow task:
 
-1. Check for existing UI/UX memory and branding first: `.ui-ux-memory.md`, design tokens, CSS variables, theme files, existing components, fonts, color scheme, and representative screens.
-2. If memory exists, follow it as the design baseline. If no memory exists, create one for the application using `templates/ui-ux-memory.md` after inspecting existing branding or asking minimum high-value questions for a fresh project.
-3. Understand the product, users, goal, platform, constraints, evidence available, research confidence, and success metric.
-4. Define the UX model: audience, jobs-to-be-done, journeys, service touchpoints, information architecture, flows, states, and content.
-5. Select an appropriate top-brand design method when frontend polish is requested: Product Cinema, Utility Command Center, Editorial Brand World, Enterprise Trust Hub, Card-Based Discovery, Conversion Simplicity, Ecosystem Bundling, Premium Restraint, Playful Familiarity, or Technical Authority.
-6. Design the interface by extending the existing memory/brand baseline unless the user explicitly asks for a redesign or specific visual modification.
-7. Validate: consistency with memory, research evidence, heuristics, WCAG, platform conventions, usability risks, i18n, ethics/privacy, edge cases, technical feasibility, metrics, and brand-fit.
-8. Update the application UI/UX memory with durable design decisions before handoff.
-9. Hand off: exact specs, tokens, component behavior, copy, states, QA checklist, implementation notes.
+1. **Discovery first.** Load `references/design-discovery-protocol.md`. For any new design task, run the 6-question discovery form before producing any visual output. Lock down: surface, audience, tone, brand context, fidelity, and constraints.
+2. **Brand extraction or visual direction.** If the user provides brand assets (URL, screenshot, hex codes), run the 5-step brand extraction protocol from `references/design-discovery-protocol.md` and write a brand-spec summary. If there is no brand, load `references/visual-directions.md`, present the 5 directions, and bind the selected direction's token block as the baseline. Never invent or guess brand colors from a company name alone.
+3. **Check project memory.** Look for `.ui-ux-memory.md` at the project root. If it exists, read all 9 sections (schema defined in `references/design-system-schema.md`) before designing. If it does not exist, create it using the schema after discovery completes.
+4. Understand the product, users, goal, platform, constraints, evidence available, research confidence, and success metric.
+5. Define the UX model: audience, jobs-to-be-done, journeys, service touchpoints, information architecture, flows, states, and content.
+6. **Industry reasoning.** Load `references/industry-reasoning-rules.md`. Match the user's product to an industry category and output the full Design System Block (pattern + style + colors + typography + effects + anti-patterns) in one step before designing.
+6a. Select an appropriate top-brand design method when frontend polish is requested. Load the corresponding brand-method skill file from references/: `references/brand-method-product-cinema.md`, `references/brand-method-utility-command-center.md`, `references/brand-method-editorial-brand-world.md`, `references/brand-method-enterprise-trust-hub.md`, `references/brand-method-card-based-discovery.md`, `references/brand-method-conversion-simplicity.md`, `references/brand-method-ecosystem-bundling.md`, `references/brand-method-premium-restraint.md`, `references/brand-method-playful-familiarity.md`, or `references/brand-method-technical-authority.md`.
+6b. Apply color psychology: load `references/color-psychology-branding.md`, identify the industry and sentiment, prescribe a complete palette with contrast ratios and rationale before any visual design begins. Use `references/color-scale-system.md` to generate full 11-step (50–950) OKLch color scales for the brand primary, neutral, and all semantic colors.
+6c. Select a UI style from `references/ui-styles-catalog.md`. Apply the style's token overrides on top of the visual direction baseline.
+7. **Junior designer warm-up.** For high-fidelity requests, show a wireframe sketch (grey blocks, real layout) first. Confirm direction before applying brand and polish. Skip if the user explicitly requests final output or provides a detailed reference.
+8. Design the interface by extending the established brand/direction baseline. Apply only tokens from the schema — no ad-hoc magic values. Apply stack-specific rules from `references/tech-stack-guidelines.md` for the user's chosen framework. For landing pages and marketing pages, follow the appropriate conversion pattern from `references/landing-page-patterns.md`. For dashboards, use the chart type guide in that same file.
+9. **Quality gates before any emit.** Load `references/output-quality-gates.md`. Run the 5-dimensional self-critique (philosophy/hierarchy/execution/specificity/restraint), pass all P0 hard gates, and clear the anti-slop blacklist. Fix and rescore until all dimensions ≥ 3. Do not emit output until all gates pass.
+10. Validate: consistency with memory, research evidence, heuristics, WCAG, platform conventions, usability risks, i18n, ethics/privacy, edge cases, technical feasibility, metrics, and brand-fit.
+11. Update `.ui-ux-memory.md` with any new durable design decisions before handoff.
+12. Hand off: exact specs, tokens, component behavior, copy, states, QA checklist, implementation notes.
 
 ## When to Use
 
@@ -456,36 +462,52 @@ Must include:
 
 ### Top-Brand Frontend Method
 
-When the user asks for a frontend that looks polished, premium, world-class, or similar to top brands, use `references/top-100-brand-website-analysis.md` before designing.
+When the user asks for a frontend that looks polished, premium, world-class, or similar to top brands, use `references/top-100-brand-website-analysis.md` before designing. Then load the dedicated skill file for the selected method.
 
 Process:
 
 1. Identify the product category and user task.
-2. Pick one primary method and one or two supporting methods:
-   - Product Cinema for premium product reveals.
-   - Utility Command Center for task-heavy ecommerce, booking, logistics, dashboards.
-   - Editorial Brand World for fashion, sports, beverage, entertainment, luxury campaigns.
-   - Enterprise Trust Hub for B2B, finance, healthcare, infrastructure, industrial.
-   - Card-Based Discovery for marketplaces, media, travel, catalogs, content libraries.
-   - Conversion Simplicity for signup, trial, app download, subscription, fintech flows.
-   - Ecosystem Bundling for multi-product suites.
-   - Premium Restraint for luxury, high-end finance, premium services.
-   - Playful Familiarity for food, toys, family, consumer brands.
-   - Technical Authority for AI, developer tools, semiconductors, infrastructure.
-3. Translate the selected methods into original tokens, layout rules, components, states, and copy.
-4. Never copy a brand exactly unless the user owns that brand. Avoid logos, proprietary images, exact trademark color combinations, and pixel-for-pixel layouts.
-5. Preserve accessibility and usability even when using cinematic, luxury, or experimental layouts.
+2. Pick one primary method and one or two supporting methods. Load the corresponding brand-method skill file for full layout, color, typography, spacing, motion, accessibility, and component specs:
+   - **Product Cinema** → `references/brand-method-product-cinema.md` — premium products, hardware, automotive, luxury.
+   - **Utility Command Center** → `references/brand-method-utility-command-center.md` — ecommerce, logistics, booking, dashboards.
+   - **Editorial Brand World** → `references/brand-method-editorial-brand-world.md` — fashion, sports, beverage, entertainment.
+   - **Enterprise Trust Hub** → `references/brand-method-enterprise-trust-hub.md` — B2B SaaS, finance, healthcare, industrial.
+   - **Card-Based Discovery** → `references/brand-method-card-based-discovery.md` — marketplaces, media, travel, catalogs.
+   - **Conversion Simplicity** → `references/brand-method-conversion-simplicity.md` — signup, trial, subscription, fintech flows.
+   - **Ecosystem Bundling** → `references/brand-method-ecosystem-bundling.md` — multi-product suites, plan comparison.
+   - **Premium Restraint** → `references/brand-method-premium-restraint.md` — ultra-luxury, high-end finance, premium services.
+   - **Playful Familiarity** → `references/brand-method-playful-familiarity.md` — food, toys, family, consumer brands.
+   - **Technical Authority** → `references/brand-method-technical-authority.md` — AI, developer tools, semiconductors, infra.
+3. Apply color psychology: load `references/color-psychology-branding.md` to validate and prescribe the palette for the business type and emotional goal. Output the full palette before any visual design.
+4. Translate the selected methods into original tokens, layout rules, components, states, and copy.
+5. Never copy a brand exactly unless the user owns that brand. Avoid logos, proprietary images, exact trademark color combinations, and pixel-for-pixel layouts.
+6. Preserve accessibility and usability even when using cinematic, luxury, or experimental layouts.
 
 Deliverable additions:
 
+- Brand method selected (with rationale).
+- Color psychology palette (business type, emotional goal, full hex palette, contrast ratios).
 - Brand-inspiration blend.
-- Selected method rationale.
 - Layout archetype.
 - Token direction.
 - Component behavior.
 - Accessibility safeguards.
 - Anti-copy/IP safety note.
 
+### Color Psychology and Branding
+
+When a user describes a product, business type, UI feel, or audience, load `references/color-psychology-branding.md` before prescribing any colors.
+
+Process:
+
+1. Identify: business category, emotional goal (trust/energy/calm/luxury/health/play/authority/innovation), positioning (mass/premium/luxury).
+2. Look up the industry prescription and sentiment mapping in the reference file.
+3. Output the complete palette: primary, secondary, accent/CTA, background, surface, text primary, text secondary, border, semantic success/warning/error/info, and dark mode variant if needed.
+4. State the psychological rationale for each choice.
+5. Warn of anti-patterns for this category.
+6. Verify all text-on-background combinations against WCAG 2.2 AA (4.5:1 body, 3:1 large text/UI).
+7. Include this palette in the `.ui-ux-memory.md` and all handoff specs.
+
 ### Frontend Implementation UX
 
 When implementing UI, first load and follow `references/ui-ux-frontend-implementation-rules.md`. Treat it as the non-negotiable frontend implementation checklist.
@@ -704,7 +726,27 @@ Use the supporting files in this skill folder when useful:
 - `references/ui-ux-frontend-implementation-rules.md` — mandatory frontend implementation rules so agents do not miss stack inspection, states, accessibility, responsiveness, tokens, QA, or handoff.
 - `references/wcag-aa-quick-reference.md` — practical WCAG 2.2 AA reference for agents.
 - `references/design-system-playbook.md` — how to create or extend a design system.
+- `references/ui-styles-catalog.md` — 20+ named UI styles (Glassmorphism, Liquid Glass, Data Brutalism, Aurora Borealis, AI-Native UI, Spatial UI, Cyberpunk, Claymorphism, Bento Grid, Soft UI Evolution, and more) with token overrides, required effects, forbidden patterns, and a style selection decision tree.
+- `references/color-scale-system.md` — 11-step (50–950) OKLch architectural color scales for all brand hue families and semantic states. Semantic role-to-step mapping and dark mode overrides. Full px+rem typography scale.
+- `references/industry-reasoning-rules.md` — one-step reasoning engine: input a product description, output a complete Design System Block (pattern, style, colors, typography, effects, anti-patterns, pre-delivery checklist) for 15+ industry categories.
+- `references/tech-stack-guidelines.md` — stack-specific component patterns, rules, and common AI mistakes for 16 frameworks: React, Next.js, Vue, Nuxt, Angular, Svelte, Astro, Remix, SolidJS, React Native, Flutter, SwiftUI, shadcn/ui, Jetpack Compose, Laravel, HTML+Tailwind.
+- `references/landing-page-patterns.md` — 12 named conversion patterns (Hero-Centric, Problem-Solution, Product Demo First, Comparison, etc.) with section-by-section structure. Plus 25-type chart selection guide with dashboard layout rules and data visualization anti-patterns.
+- `references/design-discovery-protocol.md` — 6-question discovery form, brand extraction 5-step protocol, junior designer warm-up, and discovery checklist. Load at the start of every new design task.
+- `references/output-quality-gates.md` — 5-dimensional self-critique (philosophy/hierarchy/execution/specificity/restraint), P0/P1/P2 hard gates, anti-AI-slop blacklist, and honest placeholder protocol. Run before every design emit.
+- `references/visual-directions.md` — 5 fully-specified OKLch token directions (Neutral Modern, Dark Technical, Warm Editorial, Bold Energetic, Calm Trust) for projects with no brand. Each direction includes complete CSS token block ready to bind.
+- `references/design-system-schema.md` — portable 9-section design system schema (color, typography, spacing, layout, components, motion, voice, brand, anti-patterns). Use as the template for all `.ui-ux-memory.md` files.
 - `references/top-100-brand-website-analysis.md` — top global brand website patterns and reusable frontend methods.
+- `references/brand-method-product-cinema.md` — full layout, color, typography, spacing, motion, accessibility, and component spec for the Product Cinema design method.
+- `references/brand-method-utility-command-center.md` — full spec for the Utility Command Center design method.
+- `references/brand-method-editorial-brand-world.md` — full spec for the Editorial Brand World design method.
+- `references/brand-method-enterprise-trust-hub.md` — full spec for the Enterprise Trust Hub design method.
+- `references/brand-method-card-based-discovery.md` — full spec for the Card-Based Discovery design method.
+- `references/brand-method-conversion-simplicity.md` — full spec for the Conversion Simplicity design method.
+- `references/brand-method-ecosystem-bundling.md` — full spec for the Ecosystem Bundling design method.
+- `references/brand-method-premium-restraint.md` — full spec for the Premium Restraint design method.
+- `references/brand-method-playful-familiarity.md` — full spec for the Playful Familiarity design method.
+- `references/brand-method-technical-authority.md` — full spec for the Technical Authority design method.
+- `references/color-psychology-branding.md` — color psychology, industry palette prescriptions, sentiment-to-palette mapping, contrast verification, and dark mode adaptation rules.
 - `references/ux-research-methods.md` — research planning, method selection, evidence confidence, and ethical research rules.
 - `references/usability-heuristics.md` — heuristic review, cognitive rules, and severity scoring.
 - `references/platform-guidelines.md` — web, iOS, Android/Material, Windows, desktop, kiosk, email, TV, and cross-platform rules.
@@ -716,16 +758,16 @@ Use the supporting files in this skill folder when useful:
 - `references/accessibility-advanced-patterns.md` — complex widget accessibility and screen-reader test matrix.
 - `references/ui-ux-curriculum-and-standards.md` — basic-to-advanced UI/UX curriculum and standards map.
 - `references/competitive-landscape.md` — competitor gaps and strategy for staying ahead.
-- `docs/slash-command-compatibility.md` — cross-agent `/ui-ux-master` trigger compatibility.
-- `docs/mcp-server.md` — MCP tools, resources, prompts, client configs, and smoke testing.
-- `llms.txt` — AI-readable package map for discovery by agents and indexing tools.
-- `ai-discovery/ui-ux-master.manifest.json` — machine-readable manifest for activation, entrypoints, MCP, and capabilities.
-- `system-prompts/ui-ux-master-system-add-on.md` — full system prompt add-on.
-- `system-prompts/ui-ux-master-compact.md` — compact system prompt add-on.
-- `system-prompts/ui-ux-master-mcp-add-on.md` — MCP-focused system prompt add-on.
-- `agent-templates/` — Claude, Codex, Windsurf, Antigravity, Gemini, Cursor, and universal installer templates.
-- `bin/ui-ux-master.mjs` — npm CLI installer.
-- `bin/ui-ux-master-mcp.mjs` — local read-only MCP server.
+- `docs/slash-command-compatibility.md` — cross-agent `/ui-ux-master` trigger compatibility.
+- `docs/mcp-server.md` — MCP tools, resources, prompts, client configs, and smoke testing.
+- `llms.txt` — AI-readable package map for discovery by agents and indexing tools.
+- `ai-discovery/ui-ux-master.manifest.json` — machine-readable manifest for activation, entrypoints, MCP, and capabilities.
+- `system-prompts/ui-ux-master-system-add-on.md` — full system prompt add-on.
+- `system-prompts/ui-ux-master-compact.md` — compact system prompt add-on.
+- `system-prompts/ui-ux-master-mcp-add-on.md` — MCP-focused system prompt add-on.
+- `agent-templates/` — Claude, Codex, Windsurf, Antigravity, Gemini, Cursor, and universal installer templates.
+- `bin/ui-ux-master.mjs` — npm CLI installer.
+- `bin/ui-ux-master-mcp.mjs` — local read-only MCP server.
 - `templates/ui-ux-brief.md` — intake and requirements template.
 - `templates/ui-ux-memory.md` — project UI/UX memory file template to copy into an application root as `.ui-ux-memory.md`.
 - `templates/ui-ux-audit-report.md` — audit output template.

package/docs/package-publishing.md

@@ -0,0 +1,84 @@
+# Package Publishing
+
+UI/UX Master uses npmjs as the primary public registry and GitHub as the source of truth for code, issues, releases, and automation.
+
+## Registry Strategy
+
+| Registry | Package name | Role |
+|---|---|---|
+| npmjs | `ui-ux-master` | Primary public package for `npm install`, `npx`, and discoverability |
+| GitHub Packages | `@seance1723/ui-ux-master` | Optional scoped mirror to populate the GitHub repository Packages section |
+
+Keep the npmjs package as `ui-ux-master`. If you also publish to GitHub Packages, publish a scoped mirror from CI so the source package name does not need to change.
+
+## npmjs vs GitHub Packages
+
+npmjs is the default npm registry. It is where most developers expect public JavaScript packages to live, and it is the best place for public discoverability, downloads, search, and normal `npx` usage.
+
+GitHub Packages is GitHub's package registry. It is useful for packages tied tightly to a repository or organization, private packages, and showing a package in the repository Packages section. For npm packages, GitHub Packages expects scoped package names such as `@seance1723/ui-ux-master`.
+
+The GitHub "No packages published" message does not mean the npmjs package is broken. It only means nothing has been published to GitHub Packages for that repository.
+
+## GitHub Packages Setup
+
+The root `.npmrc` intentionally configures only the `@seance1723` scope. It does not change the default registry, so normal `npm install ui-ux-master` and `npm publish` for the unscoped npmjs package still use npmjs.
+
+Scoped GitHub Packages configuration:
+
+```ini
+@seance1723:registry=https://npm.pkg.github.com
+//npm.pkg.github.com/:_authToken=${NODE_AUTH_TOKEN}
+```
+
+Local install from GitHub Packages:
+
+```bash
+npm install @seance1723/ui-ux-master --registry=https://npm.pkg.github.com
+```
+
+Many GitHub Packages installs require authentication. Use a token with `read:packages` for installs and `write:packages` for publishing.
+
+## Secrets
+
+Add these repository secrets:
+
+- `NPM_TOKEN` - npm automation token with publish access to `ui-ux-master`.
+
+GitHub Packages publishing uses the built-in `GITHUB_TOKEN` in Actions. The workflow temporarily changes the package name to `@seance1723/ui-ux-master` only for that publish job, then publishes to `https://npm.pkg.github.com/`. Do not permanently rename the root package unless you intentionally want to migrate away from the public unscoped npmjs package.
+
+## Release Strategy
+
+Use semantic versioning:
+
+- Patch: documentation fixes, metadata improvements, bug fixes that preserve behavior.
+- Minor: new agent templates, new MCP resources, new commands, backward-compatible workflow improvements.
+- Major: breaking CLI behavior, changed trigger contract, removed assets, changed package entrypoint semantics.
+
+Recommended flow:
+
+```bash
+npm run release:check
+npm version patch
+git push origin main --follow-tags
+```
+
+Create a GitHub release for the new tag, for example `v1.2.1`. The publish workflow will validate and publish.
+
+## npm Optimization Checklist
+
+- Keep `repository`, `homepage`, `bugs`, `license`, `author`, `funding`, and keywords current.
+- Keep `files` tight so generated assets and local graph/cache files stay out of the package.
+- Run `npm pack --dry-run` before publishing.
+- Prefer provenance publishing through GitHub Actions.
+- Keep README install examples short and copyable.
+- Keep `exports` stable once published.
+- Avoid runtime dependencies unless they clearly improve the CLI or MCP server.
+
+## GitHub Presentation Checklist
+
+- Add a concise About description and npm link.
+- Add SEO topics listed in README.
+- Use releases with human-written notes.
+- Keep Issues enabled and use issue templates.
+- Pin the repository on the maintainer profile.
+- Add screenshots or a short terminal GIF later if the CLI becomes more visual.

package/index.cjs

@@ -0,0 +1,39 @@
+'use strict';
+
+const path = require('node:path');
+
+const packageRoot = __dirname;
+
+const _pkg = require('./package.json');
+
+const name = 'ui-ux-master';
+const version = _pkg.version;
+const trigger = '/ui-ux-master';
+
+const bins = Object.freeze({
+  cli: path.join(packageRoot, 'bin', 'ui-ux-master.mjs'),
+  mcp: path.join(packageRoot, 'bin', 'ui-ux-master-mcp.mjs'),
+});
+
+const assets = Object.freeze({
+  skill: path.join(packageRoot, 'SKILL.md'),
+  readme: path.join(packageRoot, 'README.md'),
+  llms: path.join(packageRoot, 'llms.txt'),
+  manifest: path.join(packageRoot, 'ai-discovery', 'ui-ux-master.manifest.json'),
+});
+
+function assetPath(name) {
+  if (!Object.hasOwn(assets, name)) {
+    throw new Error(`Unknown ui-ux-master asset: ${name}`);
+  }
+  return assets[name];
+}
+
+module.exports = {
+  assetPath,
+  assets,
+  bins,
+  name,
+  trigger,
+  version,
+};

package/index.d.ts

@@ -0,0 +1,17 @@
+export declare const name = "ui-ux-master";
+export declare const version: string;
+export declare const trigger = "/ui-ux-master";
+
+export declare const bins: Readonly<{
+  cli: string;
+  mcp: string;
+}>;
+
+export declare const assets: Readonly<{
+  skill: string;
+  readme: string;
+  llms: string;
+  manifest: string;
+}>;
+
+export declare function assetPath(name: keyof typeof assets): string;

package/index.mjs

@@ -0,0 +1,30 @@
+import path from 'node:path';
+import { fileURLToPath } from 'node:url';
+import { createRequire } from 'node:module';
+
+const packageRoot = path.dirname(fileURLToPath(import.meta.url));
+const _require = createRequire(import.meta.url);
+const _pkg = _require('./package.json');
+
+export const name = 'ui-ux-master';
+export const version = _pkg.version;
+export const trigger = '/ui-ux-master';
+
+export const bins = Object.freeze({
+  cli: path.join(packageRoot, 'bin', 'ui-ux-master.mjs'),
+  mcp: path.join(packageRoot, 'bin', 'ui-ux-master-mcp.mjs'),
+});
+
+export const assets = Object.freeze({
+  skill: path.join(packageRoot, 'SKILL.md'),
+  readme: path.join(packageRoot, 'README.md'),
+  llms: path.join(packageRoot, 'llms.txt'),
+  manifest: path.join(packageRoot, 'ai-discovery', 'ui-ux-master.manifest.json'),
+});
+
+export function assetPath(name) {
+  if (!Object.hasOwn(assets, name)) {
+    throw new Error(`Unknown ui-ux-master asset: ${name}`);
+  }
+  return assets[name];
+}

package/package.json

@@ -1,17 +1,38 @@
 {
   "name": "ui-ux-master",
-  "version": "1.2.0",
-  "description": "One-stop UI/UX master skill for AI agents with opt-in /ui-ux-master activation, system prompt add-ons, AI discovery assets, and local MCP server support.",
+  "version": "1.5.0",
+  "description": "Opt-in UI/UX master skill, CLI, and MCP server for AI coding agents, with accessibility, design-system, UX research, and frontend handoff workflows.",
   "type": "module",
+  "packageManager": "npm@10.9.2",
+  "main": "./index.cjs",
+  "module": "./index.mjs",
+  "types": "./index.d.ts",
+  "exports": {
+    ".": {
+      "types": "./index.d.ts",
+      "import": "./index.mjs",
+      "require": "./index.cjs"
+    },
+    "./package.json": "./package.json",
+    "./skill": "./SKILL.md",
+    "./llms": "./llms.txt",
+    "./manifest": "./ai-discovery/ui-ux-master.manifest.json",
+    "./cli": "./bin/ui-ux-master.mjs",
+    "./mcp": "./bin/ui-ux-master-mcp.mjs"
+  },
   "bin": {
     "ui-ux-master": "./bin/ui-ux-master.mjs",
     "ui-ux-master-mcp": "./bin/ui-ux-master-mcp.mjs"
   },
+  "sideEffects": false,
   "files": [
     "SKILL.md",
     "README.md",
     "LICENSE",
     "package.json",
+    "index.cjs",
+    "index.d.ts",
+    "index.mjs",
     "references/",
     "templates/",
     "agent-templates/",
@@ -25,38 +46,92 @@
   ],
   "scripts": {
     "validate": "python scripts/validate_skill.py --release",
-    "test": "node --test tests/*.test.mjs",
+    "validate:release": "python scripts/validate_skill.py --release",
+    "build": "npm run check:syntax",
+    "check:syntax": "node --check index.mjs && node --check index.cjs && node --check bin/ui-ux-master.mjs && node --check bin/ui-ux-master-mcp.mjs",
+    "doctor": "node bin/ui-ux-master.mjs doctor --dry-run",
+    "test": "npm run test:smoke",
+    "test:smoke": "npm run check:syntax && npm run doctor",
+    "test:node": "node --test tests/*.test.mjs",
+    "ci": "npm run validate && npm run build && npm run test:smoke && npm run test:node",
+    "release:check": "npm run ci && npm run pack:dry-run",
+    "pack:dry-run": "npm pack --dry-run",
     "prepack": "npm run validate && npm test",
     "build:zip": "python scripts/build_deployment_zip.py"
   },
   "keywords": [
+    "ai",
     "ai-agent",
+    "ai-agents",
+    "ai-coding",
+    "agent",
     "ui",
     "ux",
     "ui-ux",
+    "user-experience",
+    "product-design",
+    "ux-research",
+    "frontend",
+    "frontend-development",
     "design-system",
+    "design-systems",
     "accessibility",
     "wcag",
+    "wcag-2-2",
+    "aria",
+    "a11y",
+    "mcp",
+    "model-context-protocol",
+    "mcp-server",
+    "cli",
+    "npx",
+    "prompt-engineering",
+    "system-prompt",
+    "ai-discovery",
+    "llms-txt",
     "claude-code",
+    "claude",
     "codex",
     "windsurf",
     "antigravity",
     "gemini-cli",
+    "gemini",
     "cursor",
     "slash-command",
-    "agent-skill",
-    "frontend",
-    "mcp",
-    "model-context-protocol",
-    "system-prompt",
-    "ai-discovery",
-    "llms-txt"
+    "agent-skill"
   ],
   "license": "MIT",
   "author": "Rupak Biswas",
+  "contributors": [
+    {
+      "name": "Rupak Biswas"
+    }
+  ],
+  "homepage": "https://github.com/Seance1723/UI-UX-Skills#readme",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/Seance1723/UI-UX-Skills.git"
+  },
+  "bugs": {
+    "url": "https://github.com/Seance1723/UI-UX-Skills/issues"
+  },
+  "funding": {
+    "type": "github",
+    "url": "https://github.com/sponsors/Seance1723"
+  },
   "engines": {
-    "node": ">=18"
+    "node": ">=18.18"
   },
+  "directories": {
+    "bin": "bin",
+    "doc": "docs",
+    "test": "tests"
+  },
+  "os": [
+    "darwin",
+    "linux",
+    "win32"
+  ],
   "preferGlobal": true,
   "publishConfig": {
     "access": "public"