Claude Skill

andy-p5js

AI-powered p5.js creative coding workflow - from inspiration to animation

LLM Mart · 0 points · 1 views 0 listing impressions 0 install-command copies
Download aipmandy-andy-video-skill-01-programmatic-animation_andy-p5js-27683b8.zip · 834 KB

Install

skills CLI npx skills add https://github.com/AIPMAndy/Andy-Video-Skill/tree/main/01-programmatic-animation/andy-p5js
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install aipmandy-andy-video-skill@llmmart
Git git clone https://github.com/AIPMAndy/Andy-Video-Skill.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole aipmandy/andy-video-skill collection as a plugin from our marketplace. Git is the plain clone.

README

Andy-P5JS

Andy-P5JS Logo

AI驱动的p5.js创意编码工作流 - 从灵感到动画一站式解决方案

License: MIT Version p5.js

English | 中文

🎯 What is Andy-P5JS?

Andy-P5JS is a complete AI-powered workflow for creative coding with p5.js, from inspiration discovery to final video output. It combines the best AI tools and platforms to streamline the entire creative process.

✨ Features

  • 🔍 Inspiration Discovery - Search quality templates from OpenProcessing
  • 🤖 AI Code Generation - Generate high-quality code using Gemini 3.1 Pro Canvas
  • 👁️ Live Preview - Real-time preview and debugging in p5.js Web Editor
  • ⚡ Local Optimization - Deploy locally and enhance visual effects
  • 🎨 Visual Validation - Visual regression testing with testiv/snapstack
  • 🎬 Video Output - Screen recording with Cap or video editing with ChatCut

🔄 Standard Workflow

Inspiration → AI Generation → Online Preview → Local Optimization → Visual Validation → Video Output
     ↓              ↓               ↓                  ↓                    ↓                ↓
OpenProcessing   Gemini      p5js.org         Local Server           MCP Tools       Cap/ChatCut

🚀 Quick Start

Installation

  1. Install Claude Desktop (with Andy-P5JS skill)
# Clone this repository
git clone https://github.com/andyai/andy-p5js.git
cd andy-p5js

# Copy skill to Claude skills directory
cp -r skill.md ~/.claude/skills/andy-p5js/
  1. Configure MCP Servers (optional but recommended)

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "webbridge": {
      "command": "node",
      "args": ["/path/to/webbridge/server.mjs"],
      "env": {
        "KIMI_WEBBRIDGE_URL": "http://127.0.0.1:10086"
      }
    },
    "testiv": {
      "command": "npx",
      "args": ["-y", "@testiv/mcp"]
    },
    "snapstack": {
      "command": "npx",
      "args": ["-y", "snapstack-server"]
    }
  }
}

Usage

In Claude Desktop:

/andy-p5js

Or with specific requirements:

/andy-p5js Create a 3D particle explosion effect

📖 Complete Workflow Guide

Step 1: Inspiration Search (OpenProcessing)

Visit OpenProcessing to search for quality templates.

Search Tips:

  • Use English keywords (e.g., "particle system", "3D animation")
  • Filter by likes/favorites
  • Check the license before using
  • Look for clear code structure

Step 2: AI Code Generation (Gemini Canvas)

Use the 3-step prompt engineering method:

Step 1 - Generate Style Prompt:

Generate a detailed visual style description for this p5.js animation:

Requirements: [Your requirements]

Output:
- Visual style (colors, lighting, atmosphere)
- Animation effects (particle count, motion, interaction)
- Technical details (3D/2D, physics, algorithms)

Step 2 - Generate Code with Style:

Generate complete p5.js animation code using this style:

[Prompt from Step 1]

Requirements:
- Complete HTML file (with p5.js CDN)
- Use WEBGL mode (if 3D)
- Parameterized config (CONFIG object)
- ES6 Class structure
- Responsive design
- Clear comments

Step 3 - Verify and Optimize:

  • Preview directly in Gemini Canvas
  • Adjust parameters until satisfied
  • Copy complete code

Step 3: Online Preview (p5.js Web Editor)

Visit p5.js Web Editor for real-time preview and debugging.

Advantages:

  • No local setup required
  • Real-time preview
  • Built-in examples and docs
  • Multi-file project support
  • Easy sharing and embedding

Step 4: Local Deployment & Optimization

Save code locally and start a dev server:

# Save the HTML file
# file: cosmic-dandelion.html

# Start local server
python3 -m http.server 8765

# Open in browser
open http://localhost:8765/cosmic-dandelion.html

Optimization Checklist:

  • 3D effects smooth (use WEBGL)
  • Colors harmonious (check color scheme)
  • Animation natural (use easing and noise)
  • Performance good (60fps target)
  • Responsive works (windowResized)
  • Parameters configurable (CONFIG object)

Step 5: Visual Validation (MCP Tools)

Use visual tools for quality assurance:

// testiv - Visual regression testing
testiv_capture("animation-baseline")
testiv_compare("animation-baseline", "animation-optimized")

// snapstack - Screenshot management
snapstack_capture("keyframe-1")
snapstack_timeline()

// photopea - Image post-processing
photopea_open("animation-screenshot.png")

Step 6: Video Output

Option A - Cap Screen Recording (Recommended):

  1. Open Cap app
  2. Select browser window
  3. Start recording
  4. Stop and export

Option B - ffmpeg:

# Screen record on macOS
ffmpeg -f avfoundation -i "1" -t 30 -r 60 animation.mp4

# Compress
ffmpeg -i animation.mp4 -c:v libx264 -crf 23 animation-compressed.mp4

Option C - ChatCut Video Editing:

  • Visit ChatCut
  • Upload recorded video
  • AI-assisted editing (subtitles, transitions, music)
  • Export final version

🎨 Example Projects

Cosmic Dandelion Animation

A beautiful 3D animation of birds flying away from a dandelion with particle effects.

Features:

  • 500 birds with Fibonacci sphere distribution
  • 600 twinkling stars background
  • 100 atmospheric particles
  • 15-frame trail system
  • Multi-layer glow effects
  • Perlin noise flight paths

Files:

Preview:

Cosmic Dandelion

🎬 Watch Video

🛠️ Technical Highlights

Fibonacci Sphere Distribution

// Evenly distribute points on a sphere
const phi = acos(1 - 2 * (index + 0.5) / total);
const theta = PI * (1 + sqrt(5)) * (index + 0.5);

const position = createVector(
  sin(phi) * cos(theta),
  sin(phi) * sin(theta),
  cos(phi)
);

Multi-Layer Glow Effects

// Render 3-5 layers of gradient glow
for (let r = 5; r >= 1; r--) {
  let alpha = map(r, 5, 1, 20, 150);
  fill(glowColor + alpha.toString(16).padStart(2, '0'));
  sphere(radius * (1 + r * 0.15));
}

Perlin Noise Natural Motion

// Generate natural floating motion with noise
let noise = createVector(
  map(noise(pos.x * 0.01, time * 0.1), 0, 1, -0.5, 0.5),
  map(noise(pos.y * 0.01, time * 0.1), 0, 1, -0.5, 0.5),
  map(noise(pos.z * 0.01, time * 0.1), 0, 1, -0.5, 0.5)
);
velocity.add(noise);

Trail System

// Record position history
this.trail.push(this.pos.copy());
if (this.trail.length > CONFIG.trailLength) {
  this.trail.shift();
}

// Render fading trail
for (let i = 0; i < this.trail.length - 1; i++) {
  let alpha = map(i, 0, this.trail.length, 0, 100);
  stroke(color + alpha.toString(16).padStart(2, '0'));
  line(trail[i].x, trail[i].y, trail[i].z,
       trail[i+1].x, trail[i+1].y, trail[i+1].z);
}

📊 Quality Standards

Visual Quality

  • ✅ Cosmic-level beauty (depth, lighting, effects)
  • ✅ Harmonious colors (theme-appropriate palette)
  • ✅ Smooth animation (60fps, no stuttering)
  • ✅ Rich details (particles, trails, glows)

Code Quality

  • ✅ Clear structure (Class-based architecture)
  • ✅ Parameterized config (easy to adjust)
  • ✅ Complete comments (key algorithms explained)
  • ✅ Responsive design (adapts to screen size)

Performance Standards

  • ✅ Initial load < 2s
  • ✅ Stable frame rate ≥ 60fps
  • ✅ Memory usage < 200MB
  • ✅ CPU usage < 50%

🔗 Resources

💡 Best Practices

  1. Search First, Create Second - Learn from excellent works
  2. AI-Assisted, Not Replaced - Use AI for initial version, manually refine
  3. Online Debug, Local Deploy - Use online tools for rapid iteration
  4. Parameterize Everything - Centralize all adjustable parameters in CONFIG
  5. Performance First - 60fps is the baseline
  6. Visual Validation Essential - Screenshot comparison for quantifiable improvements

🐛 Troubleshooting

Q: Gemini-generated code throws errors? A: Check p5.js version, ensure 1.7.0+, WEBGL mode requires full 3D API support.

Q: Animation stuttering? A: Reduce particle count, lower draw complexity, use object pooling for memory optimization.

Q: How to export high-quality video? A: Use Cap to record original video (60fps), then use ChatCut for lossless editing.

Q: Can I use OpenProcessing code directly? A: Most can be used, but check the license. Recommended as reference for learning.

📈 Roadmap

  • Support Three.js advanced 3D rendering
  • Integrate Tone.js audio visualization
  • Support WebGL Shader customization
  • Integrate ml5.js machine learning interaction
  • Support VR/AR output formats

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/AmazingFeature)
  3. Commit your changes (git commit -m 'Add some AmazingFeature')
  4. Push to the branch (git push origin feature/AmazingFeature)
  5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

👨‍💻 Author

Andy AIPM

🙏 Acknowledgments


⭐ Star this repo if you find it helpful!

Made with ❤️ by Andy AIPM

Skill manifest

Andy-P5JS

AI驱动的p5.js创意编码工作流 - 从灵感到动画一站式解决方案。

🎯 核心能力

  1. 灵感发现 - 从 OpenProcessing 搜索优质模板
  2. 代码生成 - 使用 Gemini 3.1 Pro Canvas 生成高质量代码
  3. 实时预览 - p5.js Web Editor 在线预览和调试
  4. 本地优化 - 本地部署和视觉效果增强
  5. 视觉验证 - 使用 testiv/snapstack 进行视觉回归测试
  6. 视频输出 - Cap 录屏或 ChatCut 剪辑

🔄 标准工作流

灵感搜索 → AI生成 → 在线预览 → 本地优化 → 视觉验证 → 视频输出
   ↓           ↓          ↓          ↓          ↓          ↓
OpenProcessing  Gemini   p5js.org   本地服务器   MCP工具    Cap/ChatCut

📋 使用场景

  • 创意动画快速原型
  • 生成艺术作品
  • 交互式可视化
  • 教学演示动画
  • 社交媒体视觉内容

🎨 示例项目

  • 宇宙级蒲公英飞鸟动画
  • 3D粒子效果系统
  • Fibonacci球体分布算法
  • 多层辉光渲染

⚙️ 配置要求

  • webbridge MCP 服务器(必需)
  • testiv/snapstack(可选,用于视觉测试)
  • photopea(可选,用于图像后处理)
  • rrweb-stack(可选,用于交互录制)
  • Cap 或 ffmpeg(可选,用于视频录制)

🚀 快速开始

# 调用技能
/andy-p5js

# 或指定需求
/andy-p5js 创建一个3D粒子爆炸效果

实现步骤

Step 1: 灵感搜索(OpenProcessing)

使用 webbridge 访问 https://openprocessing.org/ 搜索相关主题的优质模板。

搜索策略:

  • 使用英文关键词(如 "particle system", "3D animation", "generative art")
  • 筛选高点赞数/收藏数的作品
  • 优先选择有清晰代码结构的项目
  • 注意查看开源协议

webbridge 操作:

// 1. 访问 OpenProcessing
navigate("https://openprocessing.org/")

// 2. 搜索相关主题
fill("input[type='search']", "particle system 3D")
send_keys("Enter")

// 3. 截图保存灵感
screenshot("openprocessing-inspiration.png")

Step 2: AI代码生成(Gemini Canvas)

三步提示词工程法:

第一步 - 生成动画风格提示词:

请为以下p5.js动画需求生成一个详细的视觉风格描述提示词:

需求:[用户需求]

输出格式:
- 视觉风格(色彩、光影、氛围)
- 动画效果(粒子数量、运动轨迹、交互方式)
- 技术要点(3D/2D、物理引擎、算法选择)

第二步 - 使用生成的提示词生成代码:

使用以下风格描述,生成完整的p5.js动画代码:

[第一步生成的提示词]

要求:
- 完整的HTML文件(包含p5.js CDN)
- 使用WEBGL模式(如果是3D)
- 参数化配置(CONFIG对象)
- ES6 Class结构
- 响应式设计
- 注释清晰

第三步 - 验证和优化:

  • 在 Gemini Canvas 中直接预览生成的代码
  • 调整参数直到满意
  • 复制完整代码

webbridge 操作:

// 访问 Gemini Canvas
navigate("https://gemini.google.com/app?hl=en_GB")

// 选择 Gemini 3.1 Pro
click("button[aria-label='Open mode picker']")
// 选择 Pro 模式

// 输入提示词
fill("textarea[aria-label='Enter a prompt']", promptText)
send_keys("Enter")

// 等待生成并截图
screenshot("gemini-generated-code.png")

Step 3: 在线预览(p5.js Web Editor)

访问 https://editor.p5js.org/ 进行在线预览和快速迭代。

优势:

  • 无需本地环境配置
  • 实时预览和调试
  • 内置示例和文档
  • 支持多文件项目
  • 可分享和嵌入

webbridge 操作:

// 1. 访问 p5.js Editor
navigate("https://editor.p5js.org/")

// 2. 粘贴代码(需要拆分为 sketch.js)
// 注意:将HTML中的<script>部分提取到 sketch.js

// 3. 点击运行
click("button[aria-label='Play']")

// 4. 截图预览
screenshot("p5js-editor-preview.png")

Step 4: 本地部署优化

将代码保存到本地并启动开发服务器。

操作流程:

# 1. 保存文件
# 使用 Write 工具保存为 .html

# 2. 启动本地服务器
python3 -m http.server 8765 &

# 3. 使用 webbridge 打开
navigate("http://localhost:8765/animation.html")

# 4. 视觉优化
# - 调整参数(粒子数量、颜色、速度)
# - 增强效果(多层辉光、拖尾、粒子系统)
# - 性能优化(降低复杂度或优化算法)

优化检查清单:

  • 3D效果是否流畅(使用WEBGL)
  • 色彩是否和谐(检查配色方案)
  • 动画是否自然(使用缓动和噪声)
  • 性能是否达标(60fps目标)
  • 响应式是否正常(windowResized)
  • 参数是否可配置(CONFIG对象)

Step 5: 视觉验证(MCP工具)

使用视觉工具进行质量验证。

testiv - 视觉回归测试:

// 建立基线
testiv_capture("animation-baseline")

// 修改后对比
testiv_compare("animation-baseline", "animation-optimized")

snapstack - 截图管理:

// 捕获关键帧
snapstack_capture("keyframe-1")
snapstack_capture("keyframe-2")

// 生成时间轴
snapstack_timeline()

photopea - 图像后处理:

// 增强截图效果
photopea_open("animation-screenshot.png")
// 手动调整后保存

Step 6: 视频输出

方案A - Cap录屏(推荐):

# 使用 Mac 上的 Cap 应用
# 1. 打开 Cap
# 2. 选择浏览器窗口
# 3. 开始录制
# 4. 停止并导出

方案B - ffmpeg录屏:

# macOS 屏幕录制
ffmpeg -f avfoundation -i "1" -t 30 -r 60 animation.mp4

# 压缩优化
ffmpeg -i animation.mp4 -c:v libx264 -crf 23 -preset medium animation-compressed.mp4

方案C - rrweb-stack交互录制:

// 录制用户交互
rrweb_start_recording()

// 执行操作...

// 停止并保存
rrweb_stop_recording("animation-interactive.json")

视频剪辑 - ChatCut:

  • 访问 https://app.chatcut.io/
  • 上传录制的视频
  • AI辅助剪辑(自动字幕、转场、配乐)
  • 导出最终版本

📊 质量标准

视觉质量

  • ✅ 宇宙级美感(多层次、深度感、光影效果)
  • ✅ 色彩和谐(配色方案符合主题)
  • ✅ 动画流畅(60fps,无卡顿)
  • ✅ 细节丰富(粒子、拖尾、辉光)

代码质量

  • ✅ 结构清晰(Class-based架构)
  • ✅ 参数化配置(易于调整)
  • ✅ 注释完整(关键算法说明)
  • ✅ 响应式设计(适配不同屏幕)

性能标准

  • ✅ 初始加载 < 2秒
  • ✅ 帧率稳定 ≥ 60fps
  • ✅ 内存占用 < 200MB
  • ✅ CPU使用率 < 50%

🎓 技术要点

Fibonacci球体分布算法

// 在球面上均匀分布点
const phi = acos(1 - 2 * (index + 0.5) / total);
const theta = PI * (1 + sqrt(5)) * (index + 0.5);

const position = createVector(
  sin(phi) * cos(theta),
  sin(phi) * sin(theta),
  cos(phi)
);

多层辉光效果

// 渲染3-5层渐变光晕
for (let r = 5; r >= 1; r--) {
  let alpha = map(r, 5, 1, 20, 150);
  fill(glowColor + alpha.toString(16).padStart(2, '0'));
  sphere(radius * (1 + r * 0.15));
}

Perlin噪声自然运动

// 使用噪声生成自然飘动
let noise = createVector(
  map(noise(pos.x * 0.01, time * 0.1), 0, 1, -0.5, 0.5),
  map(noise(pos.y * 0.01, time * 0.1), 0, 1, -0.5, 0.5),
  map(noise(pos.z * 0.01, time * 0.1), 0, 1, -0.5, 0.5)
);
velocity.add(noise);

轨迹拖尾系统

// 记录历史位置
this.trail.push(this.pos.copy());
if (this.trail.length > CONFIG.trailLength) {
  this.trail.shift();
}

// 渲染渐变拖尾
for (let i = 0; i < this.trail.length - 1; i++) {
  let alpha = map(i, 0, this.trail.length, 0, 100);
  stroke(color + alpha.toString(16).padStart(2, '0'));
  line(trail[i].x, trail[i].y, trail[i].z,
       trail[i+1].x, trail[i+1].y, trail[i+1].z);
}

🔗 资源链接

💡 最佳实践

  1. 先搜索后创造 - 站在巨人的肩膀上,从优秀作品中学习
  2. AI辅助而非替代 - 用AI生成初版,人工精调优化
  3. 在线调试本地部署 - 利用在线工具快速迭代,本地环境深度优化
  4. 参数化一切 - CONFIG对象集中管理所有可调参数
  5. 性能优先 - 60fps是底线,牺牲细节保证流畅度
  6. 视觉验证必不可少 - 截图对比,确保优化效果可量化

🐛 常见问题

Q: Gemini生成的代码运行报错? A: 检查p5.js版本,确保使用1.7.0+,WEBGL模式需要完整的3D API支持。

Q: 动画卡顿怎么办? A: 减少粒子数量、降低绘制复杂度、使用对象池优化内存分配。

Q: 如何导出高质量视频? A: 使用Cap录制原始视频(60fps),然后用ChatCut进行无损剪辑。

Q: OpenProcessing的代码能直接用吗? A: 大部分可以,但需要注意开源协议,建议作为参考学习而非直接复制。

📈 未来扩展

  • 支持Three.js高级3D渲染
  • 集成Tone.js音频可视化
  • 支持WebGL Shader自定义
  • 集成ml5.js机器学习交互
  • 支持VR/AR输出格式

License: MIT
Repository: https://github.com/andyai/andy-p5js (待创建)
Author: Andy AIPM
Version: 1.0.0

Files (andy-video-skill)
  • assets
    • cosmic-dandelion.png 80.8 KB · in bundle
    • visual-test-suite.png 341.4 KB · in bundle
    • workflow-demo.png 434.8 KB · in bundle
  • examples
    • cosmic-dandelion-3d.html 6.5 KB · in bundle
    • cosmic-dandelion-enhanced.html 7.6 KB · in bundle
    • visual-test-suite.html 6.6 KB · in bundle
  • .gitignore 469 B · in bundle
  • LICENSE 1 KB · in bundle
  • package.json 844 B
    Loading…
  • PROJECT_SUMMARY.md 8.1 KB
    Loading…
  • README.md 10.5 KB
    # Andy-P5JS
    
    <div align="center">
    
    ![Andy-P5JS Logo](./assets/logo.png)
    
    **AI驱动的p5.js创意编码工作流 - 从灵感到动画一站式解决方案**
    
    [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
    [![Version](https://img.shields.io/badge/version-1.0.0-blue.svg)](https://github.com/andyai/andy-p5js/releases)
    [![p5.js](https://img.shields.io/badge/p5.js-1.7.0+-red.svg)](https://p5js.org/)
    
    [English](./README.md) | [中文](./README_CN.md)
    
    </div>
    
    ## 🎯 What is Andy-P5JS?
    
    Andy-P5JS is a complete AI-powered workflow for creative coding with p5.js, from inspiration discovery to final video output. It combines the best AI tools and platforms to streamline the entire creative process.
    
    ## ✨ Features
    
    - 🔍 **Inspiration Discovery** - Search quality templates from OpenProcessing
    - 🤖 **AI Code Generation** - Generate high-quality code using Gemini 3.1 Pro Canvas
    - 👁️ **Live Preview** - Real-time preview and debugging in p5.js Web Editor
    - ⚡ **Local Optimization** - Deploy locally and enhance visual effects
    - 🎨 **Visual Validation** - Visual regression testing with testiv/snapstack
    - 🎬 **Video Output** - Screen recording with Cap or video editing with ChatCut
    
    ## 🔄 Standard Workflow
    
    ```
    Inspiration → AI Generation → Online Preview → Local Optimization → Visual Validation → Video Output
         ↓              ↓               ↓                  ↓                    ↓                ↓
    OpenProcessing   Gemini      p5js.org         Local Server           MCP Tools       Cap/ChatCut
    ```
    
    ## 🚀 Quick Start
    
    ### Installation
    
    1. **Install Claude Desktop** (with Andy-P5JS skill)
    
    ```bash
    # Clone this repository
    git clone https://github.com/andyai/andy-p5js.git
    cd andy-p5js
    
    # Copy skill to Claude skills directory
    cp -r skill.md ~/.claude/skills/andy-p5js/
    ```
    
    2. **Configure MCP Servers** (optional but recommended)
    
    Add to your `claude_desktop_config.json`:
    
    ```json
    {
      "mcpServers": {
        "webbridge": {
          "command": "node",
          "args": ["/path/to/webbridge/server.mjs"],
          "env": {
            "KIMI_WEBBRIDGE_URL": "http://127.0.0.1:10086"
          }
        },
        "testiv": {
          "command": "npx",
          "args": ["-y", "@testiv/mcp"]
        },
        "snapstack": {
          "command": "npx",
          "args": ["-y", "snapstack-server"]
        }
      }
    }
    ```
    
    ### Usage
    
    **In Claude Desktop:**
    
    ```
    /andy-p5js
    ```
    
    Or with specific requirements:
    
    ```
    /andy-p5js Create a 3D particle explosion effect
    ```
    
    ## 📖 Complete Workflow Guide
    
    ### Step 1: Inspiration Search (OpenProcessing)
    
    Visit [OpenProcessing](https://openprocessing.org/) to search for quality templates.
    
    **Search Tips:**
    - Use English keywords (e.g., "particle system", "3D animation")
    - Filter by likes/favorites
    - Check the license before using
    - Look for clear code structure
    
    ### Step 2: AI Code Generation (Gemini Canvas)
    
    Use the **3-step prompt engineering method**:
    
    **Step 1 - Generate Style Prompt:**
    ```
    Generate a detailed visual style description for this p5.js animation:
    
    Requirements: [Your requirements]
    
    Output:
    - Visual style (colors, lighting, atmosphere)
    - Animation effects (particle count, motion, interaction)
    - Technical details (3D/2D, physics, algorithms)
    ```
    
    **Step 2 - Generate Code with Style:**
    ```
    Generate complete p5.js animation code using this style:
    
    [Prompt from Step 1]
    
    Requirements:
    - Complete HTML file (with p5.js CDN)
    - Use WEBGL mode (if 3D)
    - Parameterized config (CONFIG object)
    - ES6 Class structure
    - Responsive design
    - Clear comments
    ```
    
    **Step 3 - Verify and Optimize:**
    - Preview directly in Gemini Canvas
    - Adjust parameters until satisfied
    - Copy complete code
    
    ### Step 3: Online Preview (p5.js Web Editor)
    
    Visit [p5.js Web Editor](https://editor.p5js.org/) for real-time preview and debugging.
    
    **Advantages:**
    - No local setup required
    - Real-time preview
    - Built-in examples and docs
    - Multi-file project support
    - Easy sharing and embedding
    
    ### Step 4: Local Deployment & Optimization
    
    Save code locally and start a dev server:
    
    ```bash
    # Save the HTML file
    # file: cosmic-dandelion.html
    
    # Start local server
    python3 -m http.server 8765
    
    # Open in browser
    open http://localhost:8765/cosmic-dandelion.html
    ```
    
    **Optimization Checklist:**
    - [ ] 3D effects smooth (use WEBGL)
    - [ ] Colors harmonious (check color scheme)
    - [ ] Animation natural (use easing and noise)
    - [ ] Performance good (60fps target)
    - [ ] Responsive works (windowResized)
    - [ ] Parameters configurable (CONFIG object)
    
    ### Step 5: Visual Validation (MCP Tools)
    
    Use visual tools for quality assurance:
    
    ```javascript
    // testiv - Visual regression testing
    testiv_capture("animation-baseline")
    testiv_compare("animation-baseline", "animation-optimized")
    
    // snapstack - Screenshot management
    snapstack_capture("keyframe-1")
    snapstack_timeline()
    
    // photopea - Image post-processing
    photopea_open("animation-screenshot.png")
    ```
    
    ### Step 6: Video Output
    
    **Option A - Cap Screen Recording (Recommended):**
    1. Open Cap app
    2. Select browser window
    3. Start recording
    4. Stop and export
    
    **Option B - ffmpeg:**
    ```bash
    # Screen record on macOS
    ffmpeg -f avfoundation -i "1" -t 30 -r 60 animation.mp4
    
    # Compress
    ffmpeg -i animation.mp4 -c:v libx264 -crf 23 animation-compressed.mp4
    ```
    
    **Option C - ChatCut Video Editing:**
    - Visit [ChatCut](https://app.chatcut.io/)
    - Upload recorded video
    - AI-assisted editing (subtitles, transitions, music)
    - Export final version
    
    ## 🎨 Example Projects
    
    ### Cosmic Dandelion Animation
    
    A beautiful 3D animation of birds flying away from a dandelion with particle effects.
    
    **Features:**
    - 500 birds with Fibonacci sphere distribution
    - 600 twinkling stars background
    - 100 atmospheric particles
    - 15-frame trail system
    - Multi-layer glow effects
    - Perlin noise flight paths
    
    **Files:**
    - [`examples/cosmic-dandelion-3d.html`](./examples/cosmic-dandelion-3d.html) - Original version
    - [`examples/cosmic-dandelion-enhanced.html`](./examples/cosmic-dandelion-enhanced.html) - Enhanced version
    - [`examples/visual-test-suite.html`](./examples/visual-test-suite.html) - Side-by-side comparison
    
    **Preview:**
    
    ![Cosmic Dandelion](./assets/cosmic-dandelion.png)
    
    [🎬 Watch Video](./assets/cosmic-dandelion.mp4)
    
    ## 🛠️ Technical Highlights
    
    ### Fibonacci Sphere Distribution
    
    ```javascript
    // Evenly distribute points on a sphere
    const phi = acos(1 - 2 * (index + 0.5) / total);
    const theta = PI * (1 + sqrt(5)) * (index + 0.5);
    
    const position = createVector(
      sin(phi) * cos(theta),
      sin(phi) * sin(theta),
      cos(phi)
    );
    ```
    
    ### Multi-Layer Glow Effects
    
    ```javascript
    // Render 3-5 layers of gradient glow
    for (let r = 5; r >= 1; r--) {
      let alpha = map(r, 5, 1, 20, 150);
      fill(glowColor + alpha.toString(16).padStart(2, '0'));
      sphere(radius * (1 + r * 0.15));
    }
    ```
    
    ### Perlin Noise Natural Motion
    
    ```javascript
    // Generate natural floating motion with noise
    let noise = createVector(
      map(noise(pos.x * 0.01, time * 0.1), 0, 1, -0.5, 0.5),
      map(noise(pos.y * 0.01, time * 0.1), 0, 1, -0.5, 0.5),
      map(noise(pos.z * 0.01, time * 0.1), 0, 1, -0.5, 0.5)
    );
    velocity.add(noise);
    ```
    
    ### Trail System
    
    ```javascript
    // Record position history
    this.trail.push(this.pos.copy());
    if (this.trail.length > CONFIG.trailLength) {
      this.trail.shift();
    }
    
    // Render fading trail
    for (let i = 0; i < this.trail.length - 1; i++) {
      let alpha = map(i, 0, this.trail.length, 0, 100);
      stroke(color + alpha.toString(16).padStart(2, '0'));
      line(trail[i].x, trail[i].y, trail[i].z,
           trail[i+1].x, trail[i+1].y, trail[i+1].z);
    }
    ```
    
    ## 📊 Quality Standards
    
    ### Visual Quality
    - ✅ Cosmic-level beauty (depth, lighting, effects)
    - ✅ Harmonious colors (theme-appropriate palette)
    - ✅ Smooth animation (60fps, no stuttering)
    - ✅ Rich details (particles, trails, glows)
    
    ### Code Quality
    - ✅ Clear structure (Class-based architecture)
    - ✅ Parameterized config (easy to adjust)
    - ✅ Complete comments (key algorithms explained)
    - ✅ Responsive design (adapts to screen size)
    
    ### Performance Standards
    - ✅ Initial load < 2s
    - ✅ Stable frame rate ≥ 60fps
    - ✅ Memory usage < 200MB
    - ✅ CPU usage < 50%
    
    ## 🔗 Resources
    
    - **OpenProcessing**: https://openprocessing.org/
    - **p5.js Web Editor**: https://editor.p5js.org/
    - **Gemini Canvas**: https://gemini.google.com/app?hl=en_GB
    - **ChatCut**: https://app.chatcut.io/
    - **p5.js Documentation**: https://p5js.org/reference/
    - **p5.js Examples**: https://p5js.org/examples/
    
    ## 💡 Best Practices
    
    1. **Search First, Create Second** - Learn from excellent works
    2. **AI-Assisted, Not Replaced** - Use AI for initial version, manually refine
    3. **Online Debug, Local Deploy** - Use online tools for rapid iteration
    4. **Parameterize Everything** - Centralize all adjustable parameters in CONFIG
    5. **Performance First** - 60fps is the baseline
    6. **Visual Validation Essential** - Screenshot comparison for quantifiable improvements
    
    ## 🐛 Troubleshooting
    
    **Q: Gemini-generated code throws errors?**
    A: Check p5.js version, ensure 1.7.0+, WEBGL mode requires full 3D API support.
    
    **Q: Animation stuttering?**
    A: Reduce particle count, lower draw complexity, use object pooling for memory optimization.
    
    **Q: How to export high-quality video?**
    A: Use Cap to record original video (60fps), then use ChatCut for lossless editing.
    
    **Q: Can I use OpenProcessing code directly?**
    A: Most can be used, but check the license. Recommended as reference for learning.
    
    ## 📈 Roadmap
    
    - [ ] Support Three.js advanced 3D rendering
    - [ ] Integrate Tone.js audio visualization
    - [ ] Support WebGL Shader customization
    - [ ] Integrate ml5.js machine learning interaction
    - [ ] Support VR/AR output formats
    
    ## 🤝 Contributing
    
    Contributions are welcome! Please feel free to submit a Pull Request.
    
    1. Fork the repository
    2. Create your feature branch (`git checkout -b feature/AmazingFeature`)
    3. Commit your changes (`git commit -m 'Add some AmazingFeature'`)
    4. Push to the branch (`git push origin feature/AmazingFeature`)
    5. Open a Pull Request
    
    ## 📄 License
    
    This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
    
    ## 👨‍💻 Author
    
    **Andy AIPM**
    - GitHub: [@andyai](https://github.com/andyai)
    
    ## 🙏 Acknowledgments
    
    - [p5.js](https://p5js.org/) - The JavaScript library for creative coding
    - [OpenProcessing](https://openprocessing.org/) - Creative coding community
    - [Gemini](https://gemini.google.com/) - AI code generation
    - [Claude](https://claude.ai/) - AI assistant platform
    - [Kimi WebBridge](https://www.kimi.com/) - Browser automation
    
    ---
    
    <div align="center">
    
    **⭐ Star this repo if you find it helpful!**
    
    Made with ❤️ by Andy AIPM
    
    </div>
    
  • run-workflow.sh 2.7 KB
    Loading…
  • skill.md 9.6 KB
    ---
    name: andy-p5js
    version: 1.0.0
    description: AI-powered p5.js creative coding workflow - from inspiration to animation
    author: Andy
    tags: [p5js, creative-coding, animation, visual-effects, generative-art]
    dependencies: [webbridge]
    ---
    
    # Andy-P5JS
    
    AI驱动的p5.js创意编码工作流 - 从灵感到动画一站式解决方案。
    
    ## 🎯 核心能力
    
    1. **灵感发现** - 从 OpenProcessing 搜索优质模板
    2. **代码生成** - 使用 Gemini 3.1 Pro Canvas 生成高质量代码
    3. **实时预览** - p5.js Web Editor 在线预览和调试
    4. **本地优化** - 本地部署和视觉效果增强
    5. **视觉验证** - 使用 testiv/snapstack 进行视觉回归测试
    6. **视频输出** - Cap 录屏或 ChatCut 剪辑
    
    ## 🔄 标准工作流
    
    ```
    灵感搜索 → AI生成 → 在线预览 → 本地优化 → 视觉验证 → 视频输出
       ↓           ↓          ↓          ↓          ↓          ↓
    OpenProcessing  Gemini   p5js.org   本地服务器   MCP工具    Cap/ChatCut
    ```
    
    ## 📋 使用场景
    
    - 创意动画快速原型
    - 生成艺术作品
    - 交互式可视化
    - 教学演示动画
    - 社交媒体视觉内容
    
    ## 🎨 示例项目
    
    - 宇宙级蒲公英飞鸟动画
    - 3D粒子效果系统
    - Fibonacci球体分布算法
    - 多层辉光渲染
    
    ## ⚙️ 配置要求
    
    - webbridge MCP 服务器(必需)
    - testiv/snapstack(可选,用于视觉测试)
    - photopea(可选,用于图像后处理)
    - rrweb-stack(可选,用于交互录制)
    - Cap 或 ffmpeg(可选,用于视频录制)
    
    ## 🚀 快速开始
    
    ```bash
    # 调用技能
    /andy-p5js
    
    # 或指定需求
    /andy-p5js 创建一个3D粒子爆炸效果
    ```
    
    ---
    
    ## 实现步骤
    
    ### Step 1: 灵感搜索(OpenProcessing)
    
    使用 webbridge 访问 https://openprocessing.org/ 搜索相关主题的优质模板。
    
    **搜索策略:**
    - 使用英文关键词(如 "particle system", "3D animation", "generative art")
    - 筛选高点赞数/收藏数的作品
    - 优先选择有清晰代码结构的项目
    - 注意查看开源协议
    
    **webbridge 操作:**
    ```javascript
    // 1. 访问 OpenProcessing
    navigate("https://openprocessing.org/")
    
    // 2. 搜索相关主题
    fill("input[type='search']", "particle system 3D")
    send_keys("Enter")
    
    // 3. 截图保存灵感
    screenshot("openprocessing-inspiration.png")
    ```
    
    ### Step 2: AI代码生成(Gemini Canvas)
    
    **三步提示词工程法:**
    
    **第一步 - 生成动画风格提示词:**
    ```
    请为以下p5.js动画需求生成一个详细的视觉风格描述提示词:
    
    需求:[用户需求]
    
    输出格式:
    - 视觉风格(色彩、光影、氛围)
    - 动画效果(粒子数量、运动轨迹、交互方式)
    - 技术要点(3D/2D、物理引擎、算法选择)
    ```
    
    **第二步 - 使用生成的提示词生成代码:**
    ```
    使用以下风格描述,生成完整的p5.js动画代码:
    
    [第一步生成的提示词]
    
    要求:
    - 完整的HTML文件(包含p5.js CDN)
    - 使用WEBGL模式(如果是3D)
    - 参数化配置(CONFIG对象)
    - ES6 Class结构
    - 响应式设计
    - 注释清晰
    ```
    
    **第三步 - 验证和优化:**
    - 在 Gemini Canvas 中直接预览生成的代码
    - 调整参数直到满意
    - 复制完整代码
    
    **webbridge 操作:**
    ```javascript
    // 访问 Gemini Canvas
    navigate("https://gemini.google.com/app?hl=en_GB")
    
    // 选择 Gemini 3.1 Pro
    click("button[aria-label='Open mode picker']")
    // 选择 Pro 模式
    
    // 输入提示词
    fill("textarea[aria-label='Enter a prompt']", promptText)
    send_keys("Enter")
    
    // 等待生成并截图
    screenshot("gemini-generated-code.png")
    ```
    
    ### Step 3: 在线预览(p5.js Web Editor)
    
    访问 https://editor.p5js.org/ 进行在线预览和快速迭代。
    
    **优势:**
    - 无需本地环境配置
    - 实时预览和调试
    - 内置示例和文档
    - 支持多文件项目
    - 可分享和嵌入
    
    **webbridge 操作:**
    ```javascript
    // 1. 访问 p5.js Editor
    navigate("https://editor.p5js.org/")
    
    // 2. 粘贴代码(需要拆分为 sketch.js)
    // 注意:将HTML中的<script>部分提取到 sketch.js
    
    // 3. 点击运行
    click("button[aria-label='Play']")
    
    // 4. 截图预览
    screenshot("p5js-editor-preview.png")
    ```
    
    ### Step 4: 本地部署优化
    
    将代码保存到本地并启动开发服务器。
    
    **操作流程:**
    ```bash
    # 1. 保存文件
    # 使用 Write 工具保存为 .html
    
    # 2. 启动本地服务器
    python3 -m http.server 8765 &
    
    # 3. 使用 webbridge 打开
    navigate("http://localhost:8765/animation.html")
    
    # 4. 视觉优化
    # - 调整参数(粒子数量、颜色、速度)
    # - 增强效果(多层辉光、拖尾、粒子系统)
    # - 性能优化(降低复杂度或优化算法)
    ```
    
    **优化检查清单:**
    - [ ] 3D效果是否流畅(使用WEBGL)
    - [ ] 色彩是否和谐(检查配色方案)
    - [ ] 动画是否自然(使用缓动和噪声)
    - [ ] 性能是否达标(60fps目标)
    - [ ] 响应式是否正常(windowResized)
    - [ ] 参数是否可配置(CONFIG对象)
    
    ### Step 5: 视觉验证(MCP工具)
    
    使用视觉工具进行质量验证。
    
    **testiv - 视觉回归测试:**
    ```javascript
    // 建立基线
    testiv_capture("animation-baseline")
    
    // 修改后对比
    testiv_compare("animation-baseline", "animation-optimized")
    ```
    
    **snapstack - 截图管理:**
    ```javascript
    // 捕获关键帧
    snapstack_capture("keyframe-1")
    snapstack_capture("keyframe-2")
    
    // 生成时间轴
    snapstack_timeline()
    ```
    
    **photopea - 图像后处理:**
    ```javascript
    // 增强截图效果
    photopea_open("animation-screenshot.png")
    // 手动调整后保存
    ```
    
    ### Step 6: 视频输出
    
    **方案A - Cap录屏(推荐):**
    ```bash
    # 使用 Mac 上的 Cap 应用
    # 1. 打开 Cap
    # 2. 选择浏览器窗口
    # 3. 开始录制
    # 4. 停止并导出
    ```
    
    **方案B - ffmpeg录屏:**
    ```bash
    # macOS 屏幕录制
    ffmpeg -f avfoundation -i "1" -t 30 -r 60 animation.mp4
    
    # 压缩优化
    ffmpeg -i animation.mp4 -c:v libx264 -crf 23 -preset medium animation-compressed.mp4
    ```
    
    **方案C - rrweb-stack交互录制:**
    ```javascript
    // 录制用户交互
    rrweb_start_recording()
    
    // 执行操作...
    
    // 停止并保存
    rrweb_stop_recording("animation-interactive.json")
    ```
    
    **视频剪辑 - ChatCut:**
    - 访问 https://app.chatcut.io/
    - 上传录制的视频
    - AI辅助剪辑(自动字幕、转场、配乐)
    - 导出最终版本
    
    ## 📊 质量标准
    
    ### 视觉质量
    - ✅ 宇宙级美感(多层次、深度感、光影效果)
    - ✅ 色彩和谐(配色方案符合主题)
    - ✅ 动画流畅(60fps,无卡顿)
    - ✅ 细节丰富(粒子、拖尾、辉光)
    
    ### 代码质量
    - ✅ 结构清晰(Class-based架构)
    - ✅ 参数化配置(易于调整)
    - ✅ 注释完整(关键算法说明)
    - ✅ 响应式设计(适配不同屏幕)
    
    ### 性能标准
    - ✅ 初始加载 < 2秒
    - ✅ 帧率稳定 ≥ 60fps
    - ✅ 内存占用 < 200MB
    - ✅ CPU使用率 < 50%
    
    ## 🎓 技术要点
    
    ### Fibonacci球体分布算法
    ```javascript
    // 在球面上均匀分布点
    const phi = acos(1 - 2 * (index + 0.5) / total);
    const theta = PI * (1 + sqrt(5)) * (index + 0.5);
    
    const position = createVector(
      sin(phi) * cos(theta),
      sin(phi) * sin(theta),
      cos(phi)
    );
    ```
    
    ### 多层辉光效果
    ```javascript
    // 渲染3-5层渐变光晕
    for (let r = 5; r >= 1; r--) {
      let alpha = map(r, 5, 1, 20, 150);
      fill(glowColor + alpha.toString(16).padStart(2, '0'));
      sphere(radius * (1 + r * 0.15));
    }
    ```
    
    ### Perlin噪声自然运动
    ```javascript
    // 使用噪声生成自然飘动
    let noise = createVector(
      map(noise(pos.x * 0.01, time * 0.1), 0, 1, -0.5, 0.5),
      map(noise(pos.y * 0.01, time * 0.1), 0, 1, -0.5, 0.5),
      map(noise(pos.z * 0.01, time * 0.1), 0, 1, -0.5, 0.5)
    );
    velocity.add(noise);
    ```
    
    ### 轨迹拖尾系统
    ```javascript
    // 记录历史位置
    this.trail.push(this.pos.copy());
    if (this.trail.length > CONFIG.trailLength) {
      this.trail.shift();
    }
    
    // 渲染渐变拖尾
    for (let i = 0; i < this.trail.length - 1; i++) {
      let alpha = map(i, 0, this.trail.length, 0, 100);
      stroke(color + alpha.toString(16).padStart(2, '0'));
      line(trail[i].x, trail[i].y, trail[i].z,
           trail[i+1].x, trail[i+1].y, trail[i+1].z);
    }
    ```
    
    ## 🔗 资源链接
    
    - **OpenProcessing**: https://openprocessing.org/
    - **p5.js Web Editor**: https://editor.p5js.org/
    - **Gemini Canvas**: https://gemini.google.com/app?hl=en_GB
    - **ChatCut**: https://app.chatcut.io/
    - **p5.js 文档**: https://p5js.org/reference/
    - **p5.js 示例**: https://p5js.org/examples/
    
    ## 💡 最佳实践
    
    1. **先搜索后创造** - 站在巨人的肩膀上,从优秀作品中学习
    2. **AI辅助而非替代** - 用AI生成初版,人工精调优化
    3. **在线调试本地部署** - 利用在线工具快速迭代,本地环境深度优化
    4. **参数化一切** - CONFIG对象集中管理所有可调参数
    5. **性能优先** - 60fps是底线,牺牲细节保证流畅度
    6. **视觉验证必不可少** - 截图对比,确保优化效果可量化
    
    ## 🐛 常见问题
    
    **Q: Gemini生成的代码运行报错?**
    A: 检查p5.js版本,确保使用1.7.0+,WEBGL模式需要完整的3D API支持。
    
    **Q: 动画卡顿怎么办?**
    A: 减少粒子数量、降低绘制复杂度、使用对象池优化内存分配。
    
    **Q: 如何导出高质量视频?**
    A: 使用Cap录制原始视频(60fps),然后用ChatCut进行无损剪辑。
    
    **Q: OpenProcessing的代码能直接用吗?**
    A: 大部分可以,但需要注意开源协议,建议作为参考学习而非直接复制。
    
    ## 📈 未来扩展
    
    - [ ] 支持Three.js高级3D渲染
    - [ ] 集成Tone.js音频可视化
    - [ ] 支持WebGL Shader自定义
    - [ ] 集成ml5.js机器学习交互
    - [ ] 支持VR/AR输出格式
    
    ---
    
    **License**: MIT  
    **Repository**: https://github.com/andyai/andy-p5js (待创建)  
    **Author**: Andy AIPM  
    **Version**: 1.0.0
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.