planning-frontend-design-orchestrator — community planning-frontend-design-orchestrator, MockTrialGame, community, ide skills, Claude Code, Cursor, Windsurf

v1.0.0

关于此技能

适用于需要多样式概念生成和严格通行隔离以及最大多样性的前端开发代理。 Orchestrate frontend concept generation across 5 style families with up to 4 passes each, dispatching isolated Claude Code Task agents per pass with deliberate interaction design, optional media assets, and CDN libraries, then running uniqueness and visual validation.

pelchers pelchers
[0]
[0]
更新于: 3/11/2026

Killer-Skills Review

Decision support comes first. Repository text comes second.

Reference-Only Page Review Score: 6/11

This page remains useful for operators, but Killer-Skills treats it as reference material instead of a primary organic landing page.

Original recommendation layer Concrete use-case guidance Explicit limitations and caution
Review Score
6/11
Quality Score
36
Canonical Locale
en
Detected Body Locale
en

适用于需要多样式概念生成和严格通行隔离以及最大多样性的前端开发代理。 Orchestrate frontend concept generation across 5 style families with up to 4 passes each, dispatching isolated Claude Code Task agents per pass with deliberate interaction design, optional media assets, and CDN libraries, then running uniqueness and visual validation.

核心价值

赋予代理使用JSON配置文件(如style-config.json)生成多样式前端概念的能力,并利用布局唯一性目录来确保最大设计多样性,同时利用可用的库来增强功能。

适用 Agent 类型

适用于需要多样式概念生成和严格通行隔离以及最大多样性的前端开发代理。

赋予的主要能力 · planning-frontend-design-orchestrator

生成具有严格通行隔离的多样式前端设计概念
使用layout-uniqueness-catalog.json创建唯一的结构布局配置文件
使用每次传递覆盖来定制样式家族以实现定制的设计输出

! 使用限制与门槛

  • 需要访问.claude/skills/planning-frontend-design-orchestrator/references目录
  • 依赖于特定的JSON文件格式,例如style-config.json和layout-uniqueness-catalog.json

Why this page is reference-only

  • - Current locale does not satisfy the locale-governance contract.
  • - The underlying skill quality score is below the review floor.

Source Boundary

The section below is supporting source material from the upstream repository. Use the Killer-Skills review above as the primary decision layer.

实验室 Demo

Browser Sandbox Environment

⚡️ Ready to unleash?

Experience this Agent in a zero-setup browser environment powered by WebContainers. No installation required.

Boot Container Sandbox

常见问题与安装步骤

以下问题与步骤与页面结构化数据保持一致,便于搜索引擎理解页面内容。

? FAQ

planning-frontend-design-orchestrator 是什么?

适用于需要多样式概念生成和严格通行隔离以及最大多样性的前端开发代理。 Orchestrate frontend concept generation across 5 style families with up to 4 passes each, dispatching isolated Claude Code Task agents per pass with deliberate interaction design, optional media assets, and CDN libraries, then running uniqueness and visual validation.

如何安装 planning-frontend-design-orchestrator?

运行命令:npx killer-skills add pelchers/MockTrialGame/planning-frontend-design-orchestrator。支持 Cursor、Windsurf、VS Code、Claude Code 等 19+ IDE/Agent。

planning-frontend-design-orchestrator 适用于哪些场景?

典型场景包括:生成具有严格通行隔离的多样式前端设计概念、使用layout-uniqueness-catalog.json创建唯一的结构布局配置文件、使用每次传递覆盖来定制样式家族以实现定制的设计输出。

planning-frontend-design-orchestrator 支持哪些 IDE 或 Agent?

该技能兼容 Cursor, Windsurf, VS Code, Trae, Claude Code, OpenClaw, Aider, Codex, OpenCode, Goose, Cline, Roo Code, Kiro, Augment Code, Continue, GitHub Copilot, Sourcegraph Cody, and Amazon Q Developer。可使用 Killer-Skills CLI 一条命令通用安装。

planning-frontend-design-orchestrator 有哪些限制?

需要访问.claude/skills/planning-frontend-design-orchestrator/references目录;依赖于特定的JSON文件格式,例如style-config.json和layout-uniqueness-catalog.json。

安装步骤

  1. 1. 打开终端

    在你的项目目录中打开终端或命令行。

  2. 2. 执行安装命令

    运行:npx killer-skills add pelchers/MockTrialGame/planning-frontend-design-orchestrator。CLI 会自动识别 IDE 或 AI Agent 并完成配置。

  3. 3. 开始使用技能

    planning-frontend-design-orchestrator 已启用,可立即在当前项目中调用。

! 参考页模式

此页面仍可作为安装与查阅参考,但 Killer-Skills 不再把它视为主要可索引落地页。请优先阅读上方评审结论,再决定是否继续查看上游仓库说明。

Imported Repository Instructions

The section below is supporting source material from the upstream repository. Use the Killer-Skills review above as the primary decision layer.

Supporting Evidence

planning-frontend-design-orchestrator

安装 planning-frontend-design-orchestrator,这是一款面向AI agent workflows and automation的 AI Agent Skill。支持 Claude Code、Cursor、Windsurf,一键安装。

SKILL.md
Readonly
Imported Repository Instructions
The section below is supporting source material from the upstream repository. Use the Killer-Skills review above as the primary decision layer.
Supporting Evidence

Planning Frontend Design Orchestrator

Use this skill to run multi-style frontend concept ideation with strict pass isolation and maximum variety.

Config Source

  • .claude/skills/planning-frontend-design-orchestrator/references/style-config.json

Reference Catalogs

  • references/layout-uniqueness-catalog.json - 20 structural layout profiles
  • references/style-config.json - Style families with per-pass overrides
  • .claude/skills/frontend-design-subagent/references/available-libraries.json - CDN library catalog
  • .claude/skills/frontend-design-subagent/references/asset-sources.json - Approved media download sources
  • .claude/skills/frontend-design-subagent/references/external-inspiration-catalog.json - Per-pass specific inspiration URLs
  • .claude/skills/frontend-design-subagent/references/product-context.md - Product data models, terminology, and content vocabulary

Workflow

  1. Read the style config to get all style families and their pass definitions.
  2. Read the uniqueness catalog, inspiration catalog, library catalog, and asset sources.
  3. For each (style, pass) job, build a comprehensive creative brief that includes: a. Select a uniqueness profile (ensuring no two passes share the same profile). b. Apply per-pass paletteOverrides and typographyOverrides from the config. c. Include the pass's interactionProfile with all category selections and prompts. d. Include the pass's contentPersona for visual/tonal flavor. e. Include the product context from references/product-context.md — this defines all mock content (data models, terminology, view content, sample data). Content persona controls HOW it looks; product context controls WHAT it says. f. Include the pass's viewHints for per-view component directives. g. Include the pass's antiRepeat list as hard constraints. h. Include the specific inspiration references (not generic Awwwards categories). i. Include the available libraries catalog so the agent can choose 0-5 libraries. j. Include the asset sources catalog so the agent can optionally download media. k. Include the Professional Quality Standards checklist (see subagent SKILL.md). l. Dispatch as a Claude Code Task agent with subagent_type=general-purpose.
  4. The Task agent generates ALL files from scratch (no template scripts).
  5. After each pass completes, run scripts/validate-concepts-playwright.mjs --style <style> --pass <n> to capture desktop + mobile screenshots.
  6. Visual Quality Review (REQUIRED — see details below).
  7. After all passes complete, run scripts/validate-design-uniqueness.mjs for pairwise checks.
  8. Write summary index for review.

Visual Quality Review (Step 6)

After Playwright screenshots are captured for each pass, the orchestrator MUST visually inspect the screenshots and validate design quality. This is not optional — a pass that generates files and screenshots but has quality defects is NOT complete.

How to Review

For each pass, use the Read tool to view at least 4 screenshots covering different views and both viewports:

  1. validation/desktop/dashboard.png — Primary view, desktop
  2. validation/mobile/dashboard.png — Primary view, mobile
  3. validation/desktop/kanban.png — Complex layout view, desktop
  4. validation/mobile/settings.png — Form-heavy view, mobile

Quality Checklist (all must pass)

CheckWhat to Look ForFAIL Criteria
No Blank PagesEvery view must have visible, rendered contentA view is entirely white, entirely black, or shows only a nav bar with no content area
Text ContrastAll text must be readable against its backgroundLight text on light background, dark text on dark background, text obscured by background images/patterns
Mobile ResponsiveMobile views must be usable, not just shrunk desktopContent overflowing viewport, horizontal scroll required, text too small to read (<12px effective), nav completely broken or invisible
Content PopulatedViews must have real themed content, not empty shellsA view shows only headings with no cards/items/data below them, placeholder text like "Content here"
Layout IntegrityNo broken layouts or overlapping elementsElements stacked on top of each other unreadably, sidebar covering main content, footer overlapping content
Navigation VisibleUser can see how to navigate between viewsNo nav visible on desktop, no hamburger/menu accessible on mobile, nav items cut off or invisible
Visual PolishDesign should look intentional, not brokenMissing fonts falling back to Times New Roman, broken CSS showing unstyled HTML, images showing as broken icons

On Failure

If ANY check fails:

  1. Identify the specific issue (which view, which viewport, what's wrong).
  2. Read the pass's index.html, style.css, and/or app.js to diagnose the root cause.
  3. Fix the code directly (edit the files in the pass folder).
  4. Re-run Playwright screenshots for that pass.
  5. Re-review the fixed screenshots.
  6. Repeat up to 2 fix cycles. If still failing after 2 attempts, write the issues to validation/quality-issues.json and flag for manual review, but continue with other passes.

Quality Report

After review, write validation/quality-review.json with this structure:

json
1{ 2 "passId": "brutalist/pass-3", 3 "reviewedAt": "2026-02-15T...", 4 "screenshotsReviewed": ["desktop/dashboard.png", "mobile/dashboard.png", "desktop/kanban.png", "mobile/settings.png"], 5 "checks": { 6 "noBlankPages": { "pass": true, "notes": "" }, 7 "textContrast": { "pass": true, "notes": "" }, 8 "mobileResponsive": { "pass": true, "notes": "" }, 9 "contentPopulated": { "pass": true, "notes": "" }, 10 "layoutIntegrity": { "pass": true, "notes": "" }, 11 "navigationVisible": { "pass": true, "notes": "" }, 12 "visualPolish": { "pass": true, "notes": "" } 13 }, 14 "overallPass": true, 15 "fixCycles": 0 16}

IMPORTANT: New Generations vs Edits

When the user asks for NEW generations:

  • Scan existing pass folders per style to find the highest pass number
  • Create NEW pass folders starting from the next number (e.g., if pass-1 and pass-2 exist, new passes are pass-3 and pass-4)
  • Each generation run produces passesPerStyle new passes (default: 2) across all styles — the config defines how many per run, NOT the total cap
  • Existing passes are preserved as-is
  • New passes must be distinct from ALL existing passes (checked by uniqueness validation)

When the user asks for EDITS:

  • Modify the specific pass folder they reference
  • Re-run Playwright screenshots after editing

Per-Pass Variety System

Each pass variant in the style config includes these differentiation tools:

paletteOverrides

Different color palette per pass. Same style DNA, different color expression. A brutalist pass can be warm paper+red OR cool gray+electric blue — both are brutalist.

typographyOverrides

Different font families per pass within the style's typographic family. Mid-century pass-1 might use Playfair Display, pass-3 might use Fraunces. Both are serif-warm but visually distinct.

interactionProfile

Deliberate interaction design choices for every touchpoint — buttons, cards, page transitions, scroll reveals, hover states, toggles, tooltips, loading states, ambient motion, and micro-feedback. Each category can have multiple selections and an optional prompt guiding the feel. No two passes should share the same interaction profile.

contentPersona

Visual tone and metaphor per pass. The persona shapes the design language (how text, components, and interactions FEEL) while the product context from product-context.md shapes the actual content displayed (labels, metrics, column names, sample data). Content must always align with the app's real PRD.

viewHints

Per-view component/layout directives that force different patterns. Dashboard might use stat cards in one pass, radial charts in another. Kanban might use vertical columns in one, horizontal swimlanes in another.

antiRepeat

Explicit ban list from prior passes. If pass-1 used a left-rail nav and stat cards, pass-2's antiRepeat says "DO NOT use: left-rail nav, stat cards".

Required Artifacts Per Pass

  • <style>/pass-<n>/index.html - Complete HTML with all 10 views
  • <style>/pass-<n>/style.css - Full CSS with responsive breakpoints
  • <style>/pass-<n>/app.js - Navigation, interactions, library init
  • <style>/pass-<n>/README.md - Concept docs, library usage, design decisions
  • <style>/pass-<n>/validation/handoff.json - Structural metadata
  • <style>/pass-<n>/validation/inspiration-crossreference.json - Inspiration mapping

Required Visual Validation Artifacts

  • <style>/pass-<n>/validation/desktop/<view>.png — Full-page screenshot per view at 1536x960
  • <style>/pass-<n>/validation/desktop/<view>_segment-N.png — Viewport-height scroll segments (auto-generated for tall pages)
  • <style>/pass-<n>/validation/mobile/<view>.png — Full-page screenshot per view at 390x844, 2x scale
  • <style>/pass-<n>/validation/mobile/<view>_segment-N.png — Viewport-height scroll segments (auto-generated for tall pages)
  • <style>/pass-<n>/validation/report.playwright.json — Structured report with viewport info and screenshot/segment counts

A pass is NOT considered complete until all view screenshots exist. The total count is dynamic (minimum 20 full-page + additional scroll segments). Run the Playwright validation script after each pass finishes generating files.

Optional Artifacts

  • <style>/pass-<n>/assets/* - Downloaded media assets (SVGs, PNGs, Lottie JSON, video)

Scripts

  • scripts/validate-concepts-playwright.mjs - Desktop + mobile screenshot capture for every view
  • scripts/validate-design-uniqueness.mjs - Pairwise pass uniqueness checking

Notes

  • Background images and media assets are OPTIONAL, not required
  • CDN libraries are OPTIONAL — pure CSS is always valid
  • Animation libraries are style-dependent, not mandated
  • The agent picks libraries based on interaction profile fit, not by default
  • Keep style config editable for easy iteration
  • passesPerStyle in config determines how many NEW passes to generate per run (default: 2) — this is per-run, not a total cap

相关技能

寻找 planning-frontend-design-orchestrator 的替代方案 (Alternative) 或可搭配使用的同类 community Skill?探索以下相关开源技能。

查看全部

openclaw-release-maintainer

Logo of openclaw
openclaw

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

333.8k
0
AI

widget-generator

Logo of f
f

为prompts.chat的信息反馈系统生成可定制的插件小部件

149.6k
0
AI

flags

Logo of vercel
vercel

React 框架

138.4k
0
浏览器

pr-review

Logo of pytorch
pytorch

Python中具有强大GPU加速的张量和动态神经网络

98.6k
0
开发者工具