Token导航 LogoToken导航

Agent Skill · 前端设计

performance-optimization(性能优化)

用于辅助前端页面、组件、样式和交互逻辑的开发与维护。它适合让 Agent 生成或审查 React、Next.js、Vue、Tailwind、CSS 等相关代码,整理组件结构,或定位布局和性能问题。使用时需要结合项目现有设计系统、路由和构建方式,避免只生成孤立片段;涉及页面改动时,应配合本地预览和构建检查确认视觉效果。

performance-optimization 是前端设计类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证为 MIT,第三方扫描已通过,按安装量在前端设计类的 10,216 个 Skill 里排第 5,013。

总安装 753周安装 32GitHub Stars 21下载 264核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:performance-optimization(性能优化)
来源仓库:https://github.com/sanky369/vibe-building-skills
仓库路径:skills/performance-optimization
安装命令:
npx skills add https://github.com/sanky369/vibe-building-skills --skill performance-optimization
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/sanky369/vibe-building-skills --skill performance-optimization

这个 Skill 是做什么的

前端设计类 Skill 关注页面结构、视觉规范和组件写法,输出通常是可运行的界面代码。

用于辅助前端页面、组件、样式和交互逻辑的开发与维护。适合生成或审查 React、Next.js、Vue、Tailwind、CSS 等相关代码。通过 npx skills add 命令从指定 GitHub 仓库安装,需结合项目现有设计系统使用。涉及页面改动时应配合本地预览和构建检查确认视觉效果。

performance-optimization 属于前端设计类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/performance-optimization

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/sanky369/vibe-building-skills --skill performa…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

performance-optimization 怎么安装?

在终端执行 npx skills add https://github.com/sanky369/vibe-building-skills --skill performance-optimi…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

performance-optimization 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

performance-optimization 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

performance-optimization 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

performance-optimization 用的人多吗?

公开数据是累计安装 753 次,近一周安装 32 次,下载 264 次,GitHub 21 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

performance-optimization 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理前端设计类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

前端设计类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
frontend-design

前端设计 · Codex、Claude

8,823,360126,072只读MIT通过
vercel-composition-patterns

Vercel 组合模式 · Codex、Claude

3,858,81625,893需要联网MIT通过
vercel-react-native-skills

Vercel React native skills 前端 · Codex、Claude

2,468,66425,878只读MIT通过
adapt

adapt 搜索 · Codex、Claude

2,027,04023,125操作浏览器未明确通过
bolder

bolder 分析 · Codex、Claude

2,004,28823,130只读未明确通过
performance-optimization

性能优化 · 本页

75321需要联网MIT通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Performance Optimization: Perceived vs Actual Latency

Overview

Perceived latency is more important than actual latency. A 3-second operation that feels instant is better than a 1-second operation that feels slow. This skill teaches you to optimize how fast your interface *feels*, not just how fast it actually is.

Core Principle: The Linear Method

The "Linear Method" popularized by Linear.com prioritizes a steady cadence of quality over frantic sprints. Applied to performance, this means:

Speed is a feature, but perceived speed is the ultimate UX.

When a user creates a task, the UI should show it as created *instantly*. The network request happens in the background. If it fails, the UI gracefully rolls back. This pattern builds trust—users know the system will handle the details.

Actual vs Perceived Latency

Actual Latency

The real time it takes for an operation to complete (measured in milliseconds).

// Actual latency: 2000ms
const startTime = performance.now();
const result = await fetch('/api/data');
const endTime = performance.now();
console.log(`Actual latency: ${endTime - startTime}ms`);

Perceived Latency

How fast the operation *feels* to the user. This is what matters.

Factors that affect perceived latency:

  • Visual feedback (does something happen immediately?)
  • Progress indication (is the user informed of progress?)
  • Anticipation (does the UI predict what's coming?)
  • Momentum (does the interface feel responsive?)

The Latency Spectrum

< 100ms: Instant

No feedback needed. The user perceives this as instant.

// Instant - no loading state needed
const handleClick = () => {
  updateLocalState();  // Instant
  // Network request in background
  fetch('/api/update').catch(() => rollback());
};

100ms - 1s: Subtle Feedback

Show a subtle indicator. A loading spinner is overkill; a slight opacity change or color shift is enough.

/* Subtle feedback for 100-1000ms operations */
.button.loading {
  opacity: 0.7;
  cursor: wait;
}

/* Or a subtle pulse */
@keyframes pulse {
  0%, 100% { opacity: 1; }
  50% { opacity: 0.7; }
}

.button.loading {
  animation: pulse 1s ease-in-out infinite;
}

1s - 10s: Skeleton or Spinner

Show a skeleton screen (if layout is known) or spinner (if layout is unknown).

<!-- Skeleton for known layout -->
<div class="card-skeleton">
  <div class="skeleton skeleton-image"></div>
  <div class="skeleton skeleton-text"></div>
  <div class="skeleton skeleton-text"></div>
</div>

<!-- Spinner for unknown layout -->
<div class="spinner"></div>

> 10s: Progress Bar

Show a progress bar with time estimate.

<div class="progress-container">
  <div class="progress-bar" style="width: 65%"></div>
  <span class="progress-text">Uploading... 2 of 3 files</span>
</div>

Optimistic UI: The Gold Standard

Optimistic UI shows changes immediately, then syncs with the server.

Pattern: Optimistic Update

// 1. Update UI immediately (optimistic)
const newItem = { id: Date.now(), text: input.value };
setItems([...items, newItem]);
clearInput();

// 2. Send to server
const response = await fetch('/api/items', {
  method: 'POST',
  body: JSON.stringify(newItem)
});

// 3. Handle failure - rollback
if (!response.ok) {
  setItems(items);  // Remove the optimistic item
  showError('Failed to save');
}

Pattern: Optimistic Delete

// 1. Remove from UI immediately
const updatedItems = items.filter(item => item.id !== id);
setItems(updatedItems);

// 2. Send delete request
const response = await fetch(`/api/items/${id}`, {
  method: 'DELETE'
});

// 3. Handle failure - restore
if (!response.ok) {
  setItems(items);  // Restore the item
  showError('Failed to delete');
}

Pattern: Optimistic Like/Vote

// 1. Update like count immediately
const newLikeCount = isLiked ? count - 1 : count + 1;
setLikeCount(newLikeCount);
setIsLiked(!isLiked);

// 2. Send to server
const response = await fetch(`/api/items/${id}/like`, {
  method: 'POST',
  body: JSON.stringify({ liked: !isLiked })
});

// 3. Handle failure - rollback
if (!response.ok) {
  setLikeCount(count);  // Restore original count
  setIsLiked(isLiked);  // Restore original state
  showError('Failed to update');
}

Perceived Performance Strategies

1. Instant Feedback

Show that something happened immediately, even if the operation isn't complete.

// Bad - Wait for server response
const handleSubmit = async () => {
  const response = await fetch('/api/submit', { body });
  showSuccess('Saved!');
};

// Good - Show feedback immediately
const handleSubmit = async () => {
  showSuccess('Saving...');  // Instant feedback
  const response = await fetch('/api/submit', { body });
  if (!response.ok) showError('Failed to save');
};

2. Progress Indication

Show progress for long operations. Even fake progress (that doesn't go to 100%) feels better than nothing.

// Fake progress for unknown duration
const startFakeProgress = () => {
  let progress = 10;
  const interval = setInterval(() => {
    progress += Math.random() * 20;
    if (progress > 90) progress = 90;  // Never reach 100%
    setProgress(progress);
  }, 500);
  return interval;
};

// Real progress for known duration
const startRealProgress = (total) => {
  let completed = 0;
  const interval = setInterval(() => {
    completed++;
    setProgress((completed / total) * 100);
    if (completed === total) clearInterval(interval);
  }, 100);
  return interval;
};

3. Anticipatory Design

Preload content before the user asks for it.

// Preload next page when user scrolls near bottom
const handleScroll = () => {
  if (isNearBottom()) {
    preloadNextPage();  // Load before user clicks
  }
};

// Preload hover targets
const handleMouseEnter = () => {
  preloadUserProfile(userId);  // Load on hover
};

4. Momentum Conservation

Respect the user's gesture velocity. If they swipe fast, the animation should feel fast.

// Respect gesture velocity
const handleSwipe = (velocity) => {
  const distance = Math.abs(velocity) * 100;  // Distance based on velocity
  const duration = Math.max(300, distance / 500);  // Duration based on distance

  animate({
    from: currentPosition,
    to: currentPosition + distance,
    duration: duration,
    easing: 'ease-out'
  });
};

Latency Masking Techniques

1. Skeleton Screens

Show a placeholder that matches the content structure.

<!-- Skeleton that matches actual content -->
<div class="card-skeleton">
  <div class="skeleton skeleton-image" style="height: 200px;"></div>
  <div class="skeleton skeleton-text" style="width: 80%;"></div>
  <div class="skeleton skeleton-text" style="width: 60%;"></div>
  <div class="skeleton skeleton-text" style="width: 40%;"></div>
</div>

Why it works: The brain constructs a spatial map while waiting. When real content arrives, it feels like a natural transition, not a sudden appearance.

2. Blur-Up Technique

Load a low-quality image first, then replace with high-quality.

<img
  src="image-low-quality.jpg"
  srcset="image-high-quality.jpg 1x"
  loading="lazy"
/>

<!-- Or with CSS -->
<div class="image-container">
  <img class="image-blur" src="image-low.jpg" />
  <img class="image-sharp" src="image-high.jpg" />
</div>
.image-blur {
  filter: blur(20px);
  opacity: 1;
  transition: opacity 300ms;
}

.image-sharp {
  opacity: 0;
  transition: opacity 300ms;
}

.image-sharp.loaded {
  opacity: 1;
}

3. Progressive Enhancement

Show basic content immediately, enhance as data loads.

<!-- Show immediately -->
<article class="post">
  <h1>Post Title</h1>
  <p>First paragraph (server-rendered)</p>
</article>

<!-- Load additional content -->
<div id="comments-container">
  <!-- Comments load via JavaScript -->
</div>

4. Staggered Animations

Animate elements in sequence to make loading feel faster.

/* Stagger animations */
.list-item {
  animation: slideIn 300ms ease-out;
  animation-fill-mode: both;
}

.list-item:nth-child(1) { animation-delay: 0ms; }
.list-item:nth-child(2) { animation-delay: 50ms; }
.list-item:nth-child(3) { animation-delay: 100ms; }
.list-item:nth-child(4) { animation-delay: 150ms; }

@keyframes slideIn {
  from {
    opacity: 0;
    transform: translateY(20px);
  }
  to {
    opacity: 1;
    transform: translateY(0);
  }
}

Why it works: Staggered animations create the illusion of faster loading. Content appears to cascade in, which feels more natural than everything appearing at once.

Network Optimization

1. Request Batching

Combine multiple requests into one.

// Bad - 5 separate requests
await fetch('/api/user');
await fetch('/api/posts');
await fetch('/api/comments');
await fetch('/api/likes');
await fetch('/api/followers');

// Good - 1 batched request
const data = await fetch('/api/batch', {
  method: 'POST',
  body: JSON.stringify({
    requests: ['user', 'posts', 'comments', 'likes', 'followers']
  })
});

2. Request Deduplication

Prevent duplicate requests for the same data.

const requestCache = new Map();

const fetchData = async (url) => {
  // Return cached promise if request is in flight
  if (requestCache.has(url)) {
    return requestCache.get(url);
  }

  const promise = fetch(url).then(r => r.json());
  requestCache.set(url, promise);

  try {
    return await promise;
  } finally {
    requestCache.delete(url);  // Clear after completion
  }
};

3. Connection Pooling

Reuse HTTP connections.

// Use HTTP/2 or HTTP/3 (automatic in modern browsers)
// Or manually manage connection pool
const connectionPool = [];
const MAX_CONNECTIONS = 6;

const getConnection = () => {
  if (connectionPool.length < MAX_CONNECTIONS) {
    return new Connection();
  }
  return connectionPool.shift();
};

Measuring Perceived Performance

Core Web Vitals

  1. LCP (Largest Contentful Paint) — When main content is visible
  2. FID (First Input Delay) — Time to respond to user input
  3. CLS (Cumulative Layout Shift) — Visual stability
// Measure LCP
const observer = new PerformanceObserver((list) => {
  const entries = list.getEntries();
  const lastEntry = entries[entries.length - 1];
  console.log('LCP:', lastEntry.renderTime || lastEntry.loadTime);
});
observer.observe({ entryTypes: ['largest-contentful-paint'] });

// Measure FID
const fidObserver = new PerformanceObserver((list) => {
  list.getEntries().forEach((entry) => {
    console.log('FID:', entry.processingDuration);
  });
});
fidObserver.observe({ entryTypes: ['first-input'] });

How to Use This Skill with Claude Code

Optimize Perceived Performance

"I'm using the performance-optimization skill. Can you help me:
- Implement optimistic UI for my form
- Add skeleton screens for loading states
- Measure Core Web Vitals
- Identify perceived latency issues"

Create Loading State Strategy

"Can you create a loading state strategy for my app?
- < 100ms: No feedback
- 100ms-1s: Subtle feedback
- 1s-10s: Skeleton screen
- > 10s: Progress bar"

Implement Staggered Animations

"Can you add staggered animations to my list?
- Animate items in sequence
- 50ms delay between items
- Smooth slide-in effect
- Respect prefers-reduced-motion"

Integration with Other Skills

  • loading-states — Loading state design
  • interaction-physics — Smooth animations
  • accessibility-excellence — Respect motion preferences
  • component-architecture — Optimistic components

Key Principles

1. Perceived > Actual How fast it feels matters more than how fast it is.

2. Instant Feedback Show something happened immediately.

3. Progress Indication Keep users informed during long operations.

4. Anticipation Preload before the user asks.

5. Momentum Respect user gesture velocity.

Checklist: Is Your Performance Ready?

  • Optimistic UI for mutations
  • Skeleton screens for 1-10s loads
  • Progress bars for >10s operations
  • Instant feedback for all actions
  • Staggered animations for lists
  • Preloading for anticipated actions
  • Request batching and deduplication
  • LCP < 2.5s
  • FID < 100ms
  • CLS < 0.1

Optimized perceived performance transforms interfaces from sluggish to snappy.

平台分布

Claude Code

31.42%
按下载量换算83

Codex

22.27%
按下载量换算59

Gemini CLI

16.13%
按下载量换算43

OpenCode

11.59%
按下载量换算31

Antigravity

7.98%
按下载量换算21

Cursor

3.15%
按下载量换算8

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。

继续浏览同类 Skills