feat: add plugin marketplace and screenshot redesign workflow

.claude-plugin/marketplace.json

@@ -0,0 +1,31 @@
+{
+  "$schema": "https://json.schemastore.org/claude-code-marketplace.json",
+  "name": "taste-skill",
+  "description": "Portable anti-slop frontend and image-generation skills for Claude Code.",
+  "owner": {
+    "name": "Leonxlnx",
+    "email": "hello@tasteskill.dev"
+  },
+  "plugins": [
+    {
+      "name": "taste-skill",
+      "displayName": "Taste Skill",
+      "description": "Anti-slop frontend and image-generation skills for premium UI design, redesigns, brand kits, and image-to-code workflows.",
+      "source": "./",
+      "author": {
+        "name": "Leonxlnx"
+      },
+      "homepage": "https://tasteskill.dev",
+      "repository": "https://github.com/Leonxlnx/taste-skill",
+      "license": "MIT",
+      "keywords": [
+        "frontend",
+        "design",
+        "skills",
+        "brandkit",
+        "image-generation"
+      ],
+      "category": "development"
+    }
+  ]
+}

.claude-plugin/plugin.json

@@ -0,0 +1,19 @@
+{
+  "$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
+  "name": "taste-skill",
+  "displayName": "Taste Skill",
+  "description": "Portable anti-slop frontend and image-generation skills for Claude Code.",
+  "author": {
+    "name": "Leonxlnx"
+  },
+  "homepage": "https://tasteskill.dev",
+  "repository": "https://github.com/Leonxlnx/taste-skill",
+  "license": "MIT",
+  "keywords": [
+    "frontend",
+    "design",
+    "skills",
+    "brandkit",
+    "image-generation"
+  ]
+}

README.md

@@ -59,6 +59,17 @@ npx skills add https://github.com/Leonxlnx/taste-skill --skill "design-taste-fro
 
 You can also copy any `SKILL.md` into your project or paste it into ChatGPT / Codex conversations.
 
+### Claude Code plugin marketplace
+
+Claude Code can install Taste Skill as a marketplace-backed plugin:
+
+```text
+/plugin marketplace add Leonxlnx/taste-skill
+/plugin install taste-skill@taste-skill
+```
+
+This keeps the existing `skills/` layout unchanged while making the skills discoverable through Claude Code's `/plugin` UI.
+
 ### Updating from the previous version
 
 The default `taste-skill` (install name `design-taste-frontend`) is now **v2 (experimental)**, a substantial rewrite of the original v1. If you already have v1 installed, just re-run the install command and you will be upgraded:
@@ -90,6 +101,7 @@ The `Install name` column is the exact value you pass to `--skill`.
 | **gpt-tasteskill** | `gpt-taste` | Stricter variant for GPT/Codex: higher layout variance, stronger GSAP direction, aggressive anti-slop. |
 | **image-to-code-skill** | `image-to-code` | Image-first pipeline: generate site references, analyze them, then implement the frontend to match. |
 | **redesign-skill** | `redesign-existing-projects` | Existing projects: audit the UI first, then fix layout, spacing, hierarchy, styling. |
+| **redesign-imagegen-codex** | `redesign-imagegen-codex` | Screenshot-first redesign workflow for Codex: capture the current UI, generate redesigned references, then implement targeted upgrades. |
 | **soft-skill** | `high-end-visual-design` | Polished, calm, expensive UI with softer contrast, whitespace, premium fonts, spring motion. |
 | **output-skill** | `full-output-enforcement` | When the model ships half-finished work: full output, no placeholder comments. |
 | **minimalist-skill** | `minimalist-ui` | Editorial product UI (Notion/Linear vibes), restrained palette, crisp structure. |
@@ -104,7 +116,7 @@ These produce design images only (no code). Use with ChatGPT Images, Codex image
 | --- | --- | --- |
 | **imagegen-frontend-web** | `imagegen-frontend-web` | Website comps: hero, landing, multi-section with strong typography, spacing, anti-slop art direction. |
 | **imagegen-frontend-mobile** | `imagegen-frontend-mobile` | Mobile screens and flows: iOS/Android/cross-platform, mockups, readable type, coherent sets. |
-| **brandkit** | `brandkit` | Brand-kit boards: logo directions, palettes, type, identity applications across categories. |
+| **brandkit** | `brandkit` | Brand-kit boards for new or existing brands: logo directions, palettes, type, identity applications across categories. |
 
 ### Which one should I use?
 
@@ -113,6 +125,7 @@ These produce design images only (no code). Use with ChatGPT Images, Codex image
 - Use **gpt-taste** when you want the stricter GPT/Codex-oriented rules and motion/layout enforcement. 
 - Use **image-to-code-skill** for image → analyze → code website workflows. 
 - Use **redesign-skill** to improve an existing codebase instead of greenfield styling. 
+- Use **redesign-imagegen-codex** when Codex can open the current UI and should redesign from screenshots before touching code.
 - Add **soft-skill**, **minimalist-skill**, or **brutalist-skill** when the visual direction is already chosen. 
 - Add **output-skill** if the agent keeps truncating output. 
 - Use **imagegen-frontend-web**, **imagegen-frontend-mobile**, or **brandkit** when the deliverable is **images** (comps, flows, identity boards), then pass results to your coding agent.

skill.sh

@@ -10,6 +10,7 @@ declare -A SKILLS=(
   [imagegen-frontend-mobile]="skills/imagegen-frontend-mobile/SKILL.md"
   [brandkit]="skills/brandkit/SKILL.md"
   [redesign-skill]="skills/redesign-skill/SKILL.md"
+  [redesign-imagegen-codex]="skills/redesign-imagegen-codex/SKILL.md"
   [soft-skill]="skills/soft-skill/SKILL.md"
   [output-skill]="skills/output-skill/SKILL.md"
   [minimalist-skill]="skills/minimalist-skill/SKILL.md"

skills/brandkit/SKILL.md

@@ -94,6 +94,52 @@ If the user gives references, match their quality and rhythm, not their exact co
 
 ---
 
+# BRAND ASSET INTAKE
+
+Before generating, determine whether this is a net-new brand or an existing brand system. Do not assume the user wants a full invention from zero.
+
+If the prompt does not clearly say whether brand assets already exist, ask one short intake question before generating:
+
+> Before I create the brand kit, do you already have existing brand assets such as a logo, color palette, typography, product UI, photography, or visual references? If yes, I will build around them. If not, I will create the brand kit from scratch.
+
+Do not ask a long questionnaire unless the user requests brand strategy help. One question is enough to choose the workflow.
+
+## Brand Kit Modes
+
+Use one of these modes:
+
+1. **Create from scratch**
+   - Use when the user has no logo, colors, typography, or brand direction.
+   - Generate the complete identity system: logo concept, palette, type direction, applications, imagery, and system details.
+   - You may invent the logo mark, symbolic metaphor, palette, and visual world.
+
+2. **Build on existing brand**
+   - Use when the user provides a logo, colors, typography, product screenshots, brand deck, website, or clear visual references.
+   - Preserve recognizable brand equity. Do not replace the logo, rename the brand, or discard existing colors without explicit permission.
+   - Turn the existing assets into a more complete brand system: usage rules, applications, palette refinement, typography hierarchy, mockups, image direction, and detail language.
+   - If an existing asset is weak, improve its presentation and surrounding system first. Only propose a redesigned alternative if the user explicitly asks for a refresh.
+
+3. **Extend partial inputs**
+   - Use when the user provides only some ingredients, such as a name and palette, a logo without type, or a mood reference without assets.
+   - Lock the provided inputs, then fill only the missing parts.
+   - Make clear internally which elements are inherited and which elements are newly generated.
+
+## Intake Signals To Preserve
+
+When assets exist, identify and preserve:
+
+- brand name spelling and casing
+- existing logo or wordmark structure
+- primary colors and any forbidden colors
+- typography or font references
+- product category and audience
+- visual references, mood, or competitive set
+- required applications, such as app icon, website, packaging, deck, social, or stationery
+
+Existing assets are source material, not decoration. The generated board should make them feel more systematic, premium, and usable.
+
+---
+
 # BRAND STRATEGY FIRST
 
 Before generating, infer the brand strategy.
@@ -740,11 +786,14 @@ Use this structure internally:
 Create a premium brand-kit overview image for "[BRAND NAME]".
 
 Brand strategy:
+- mode: [create from scratch / build on existing brand / extend partial inputs]
 - category: [category]
 - audience: [audience]
 - personality: [traits]
 - core metaphor: [metaphor]
 - logo idea: [how the mark combines symbol + name + category meaning]
+- provided assets to preserve: [logo / colors / typography / references / none]
+- missing system pieces to create: [palette / type / applications / image direction / details]
 
 Layout:
 [3×3 / 2×3 / custom] grid on a dark or light presentation canvas with strong gutters, clean alignment, and refined negative space.
@@ -775,6 +824,8 @@ readable, minimal, high hierarchy, no tiny fake text.
 Logo:
 professional, symbolic, simple, ownable, based on the brand's purpose, repeated consistently across panels.
 
+If mode is "build on existing brand", preserve the supplied logo and recognizable brand assets. Do not invent a replacement identity unless the user explicitly asks for a refresh.
+
 ---
 
 # FINAL OUTPUT STANDARD

skills/llms.txt

@@ -6,6 +6,7 @@ imagegen-frontend-web: Image-generation-only skill for creating premium website
 imagegen-frontend-mobile: Image-generation-only skill for creating premium mobile app screen concepts and flows. Does not write code.
 brandkit: Image-generation-only skill for creating premium brand-kit overview images with logo concepts, identity systems, color palettes, typography, and mockups. Does not write code.
 redesign-skill: For upgrading existing projects by auditing and fixing design problems.
+redesign-imagegen-codex: Screenshot-first redesign workflow for Codex. Capture the live UI with Chrome DevTools MCP or another browser tool, audit screenshots, generate redesigned references, then implement targeted changes.
 soft-skill: Focuses on an expensive, soft UI look with premium fonts, whitespace, depth, and smooth animations.
 output-skill: Prevents AI from being lazy, skipping code blocks, or using placeholder comments.
 minimalist-skill: Enforces clean, editorial-style interfaces (Notion/Linear style) with strict monochrome palettes.

skills/redesign-imagegen-codex/SKILL.md

@@ -0,0 +1,71 @@
+---
+name: redesign-imagegen-codex
+description: Screenshot-first redesign workflow for Codex. Use when improving an existing web UI with browser access: capture the live page with Chrome DevTools MCP or another browser tool, audit the screenshots, generate redesigned visual references, then implement targeted code changes against the existing project.
+---
+
+# Redesign Imagegen Codex
+
+Use this skill for existing websites or apps where visual quality matters and Codex can inspect the running UI in a browser.
+
+The workflow is screenshot-first:
+
+1. **Run the project**
+   - Read the repo first.
+   - Identify the framework, package manager, and dev command.
+   - Start the local app or use the user's provided URL.
+   - Do not redesign from code alone when the UI can be opened.
+
+2. **Capture the current UI**
+   - Prefer Chrome DevTools MCP when available.
+   - Otherwise use Playwright, browser screenshots, or the available screenshot tool.
+   - Capture desktop and mobile viewports.
+   - Capture key states: hero/top screen, major sections, navigation, forms, empty/error/loading states when reachable.
+   - Save screenshots as artifacts so the redesign has a concrete baseline.
+
+3. **Audit before imagining**
+   - Inspect the screenshots before generating new visuals.
+   - Identify what must be preserved: content, IA, conversion goals, brand assets, product constraints, accessibility needs.
+   - Identify visual debt: weak hierarchy, generic AI patterns, spacing issues, color/contrast problems, awkward responsive behavior, missing interaction states.
+   - State whether this is a preserve-and-polish redesign or a full visual overhaul.
+
+4. **Generate redesigned references**
+   - Generate new reference images from the captured UI and audit.
+   - Use one clear image per important section or state.
+   - Keep references implementation-friendly: readable typography, visible spacing, clear component boundaries, realistic responsive behavior.
+   - Do not generate an unrelated fantasy concept. The new references must answer the existing product's actual page structure.
+
+5. **Analyze the generated references**
+   - Extract layout, spacing, typography, color, surface treatment, motion, and component rules.
+   - Decide what can be implemented safely with the current stack.
+   - Avoid migrations, new design systems, or extra dependencies unless the existing project already supports them or the user explicitly asks.
+
+6. **Implement targeted changes**
+   - Work with the existing framework and styling system.
+   - Preserve behavior, routes, forms, data loading, and accessibility semantics.
+   - Make scoped visual upgrades instead of rebuilding the app from scratch.
+   - If generated imagery is needed but unavailable at implementation time, create clear asset slots and document the required image dimensions.
+
+7. **Verify in-browser**
+   - Reopen the modified page.
+   - Capture after screenshots for the same viewports and states.
+   - Check layout, text fit, contrast, focus states, hover/pressed states, and mobile overflow.
+   - Run the closest available tests or build command.
+
+## Rules
+
+- Do not skip the screenshot baseline.
+- Do not rely on screenshots only from one viewport.
+- Do not overwrite existing brand identity unless the user asks for a full rebrand.
+- Do not replace working product UI with static mockups.
+- Do not invent fake data where the existing app already provides real data.
+- Do not use image generation as decoration; use it to create buildable design direction.
+
+## Output Contract
+
+When reporting back, include:
+
+- baseline screenshots captured
+- redesign references generated
+- files changed
+- verification commands run
+- any visual risks or assets still needed