Token导航 LogoToken导航

Agent Skill · 前端设计

generate-readme-screenshots(generate README screenshots 控制)

用于辅助文档、README、Markdown、说明文和内容稿件的整理与改写。它适合让 Agent 提炼结构、补齐章节、统一术语、检查链接或把零散材料整理成可读文档。使用时应保留项目已有事实、命令和路径,不要把未确认的信息写成确定结论;涉及对外文案时,还需要控制语气,避免过度营销或夸大能力。

generate-readme-screenshots 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,权限为操作浏览器;许可证信息还不明确,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 5,315。

总安装 698周安装 30GitHub Stars 1,691下载 245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
unknown

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:generate-readme-screenshots(generate README screenshots 控制)
来源仓库:https://github.com/coder/mux
仓库路径:skills/generate-readme-screenshots
安装命令:
npx skills add https://github.com/coder/mux --skill generate-readme-screenshots
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。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/coder/mux --skill generate-readme-screenshots

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。

辅助 README 文档整理与改写,提升说明文可读性。适合在 Codex、Claude、Cursor、Gemini CLI 中提炼结构、统一术语或检查链接。通过 GitHub 安装,使用 npx skills add 命令添加技能。应保留项目事实,避免将未确认信息写成确定结论。

generate-readme-screenshots 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/generate-readme-screenshots

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/coder/mux --skill generate-readme-screenshots;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

    它需要操作浏览器,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

generate-readme-screenshots 怎么安装?

在终端执行 npx skills add https://github.com/coder/mux --skill generate-readme-screenshots。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

generate-readme-screenshots 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。

generate-readme-screenshots 需要什么权限?

权限级别是操作浏览器。该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

generate-readme-screenshots 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

generate-readme-screenshots 用的人多吗?

公开数据是累计安装 698 次,近一周安装 30 次,下载 245 次,GitHub 1,691 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。

generate-readme-screenshots 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。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 来源仓库有 1,691 星,维护活跃度相对可查。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 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。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 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
generate-readme-screenshots

generate README screenshots 控制 · 本页

6981,691操作浏览器未明确通过

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

Generate README Screenshots

Purpose

The README screenshots in docs/img/*.webp are not manually captured — they are deterministically generated from Storybook stories defined in src/browser/stories/App.readmeScreenshots.stories.tsx. When any of these stories change visually (detected by Chromatic or local inspection), the corresponding WebP assets must be regenerated and committed.

Quick Reference

# 1. Build Storybook
make storybook-build

# 2. Serve the static build (keep running in background)
python3 -m http.server 6006 -d storybook-static &

# 3. Capture all screenshots (3800px, WebP quality 90)
bun run scripts/capture-readme-screenshots.ts

# 4. Capture a single story
bun run scripts/capture-readme-screenshots.ts --story CodeReview

Prerequisites

The capture pipeline requires these tools. If any are missing, stop immediately and report the issue to the user with full detail — do not silently skip or produce degraded output.

DependencyPurposeCheck command
bunScript runtimebun --version
playwrightHeadless browser automationbun -e "require('playwright')"
sharpImage processing (resize, WebP)bun -e "require('sharp')"
python3Static file server for Storybook buildwhich python3

Prerequisite Verification

Before running the pipeline, verify all dependencies:

bun --version && \
bun -e "require('playwright'); require('sharp'); console.log('deps ok')" && \
which python3

If any of these fail:

  1. Do not proceed with screenshot generation.
  2. Report the exact failing command and error to the user.
  3. Suggest remediation:

- Missing playwright: bun add -d playwright && bunx playwright install chromium - Missing sharp: bun add -d sharp - Missing python3: suggest installing Python or using an alternative static server - Missing bun: this is a hard requirement — the entire repo uses bun

Architecture

Resolution & Output

ParameterValue
CSS Viewport1900 × 1188
Device Scale Factor (DPR)2
Native capture resolution3800 × 2376
Output formatWebP (quality 90, lanczos3 resampling)
Output directorydocs/img/

Story → Screenshot Mapping

Story Export NameOutput FileHas Interaction
CodeReviewcode-review.webpNo
AgentStatusSidebaragent-status.webpNo
GitStatusPopovergit-status.webpYes (open dialog + toggle mode)
PlanMermaidWithCostsplan-mermaid.webpNo
AutoModeAgentSwitchingauto-mode.webpYes (open creation card)
CostsTabRichcosts-tab.webpNo
ContextManagementDialogcontext-management.webpYes (open settings dialog)
MobileServerModemobile-server-mode.webpNo
OrchestrateAgentsorchestrate-agents.webpNo

Key Files

FileRole
src/browser/stories/App.readmeScreenshots.stories.tsxStory definitions and mock data
scripts/capture-readme-screenshots.tsPlaywright + Sharp capture pipeline
src/browser/stories/mockFactory.tsDeterministic mock data factories
src/browser/stories/mocks/orpc.tsMock ORPC backend (terminal, git, secrets)
docs/img/*.webpOutput screenshot assets

Step-by-Step Procedure

1. Build Storybook

make storybook-build

This produces storybook-static/ in the repo root.

2. Start a Static Server

The capture script expects Storybook at http://localhost:6006. Use a background bash task (not nohup or & in foreground) so the server survives:

# As a background bash task with ~10min lifetime
python3 -m http.server 6006 -d storybook-static

Verify the server is reachable:

bun -e "const r = await fetch('http://localhost:6006'); console.log(r.status)"
Important: curl may succeed when bun fetch fails (different DNS resolution). Always verify with bun -e "fetch(...)" since the capture script uses bun's fetch.

3. Run the Capture Script

bun run scripts/capture-readme-screenshots.ts --storybook-url http://localhost:6006

The script:

  • Iterates all 9 stories sequentially
  • Opens a fresh Playwright page per story
  • Waits for networkidle + 2s stabilization
  • Runs playInteraction if defined (with up to 3 retries for flaky interactions)
  • Captures full-page PNG, then converts to WebP via Sharp
  • Writes to docs/img/<outputFile>

4. Handle Failures

Stories with playInteraction (GitStatusPopover, AutoModeAgentSwitching, ContextManagementDialog) can be flaky under sequential load due to UI timing. The script retries each up to 3 times. If failures persist:

# Retry individual failed stories
bun run scripts/capture-readme-screenshots.ts --story GitStatusPopover
bun run scripts/capture-readme-screenshots.ts --story AutoModeAgentSwitching

5. Verify & Commit

# Check all 9 files are present and 3800px wide
for f in code-review agent-status git-status plan-mermaid auto-mode costs-tab context-management mobile-server-mode orchestrate-agents; do
  bun -e "const s = require('sharp'); const m = await s('docs/img/${f}.webp').metadata(); console.log('${f}:', m.width + 'x' + m.height)"
done

# Stage and commit
git add docs/img/*.webp
git commit -m "regenerate README screenshots at 3800px"

When to Run This

  1. Chromatic flags visual changes in any Docs/README Screenshots story
  2. Story data changes in App.readmeScreenshots.stories.tsx or mockFactory.ts
  3. UI component changes that affect the visual appearance of screenshot stories
  4. Explicit user request to refresh README images
  5. After rebasing on main when upstream changes affect shared UI components

Troubleshooting

"Storybook is not running"

The capture script verifies Storybook is reachable before starting. Ensure:

  • make storybook-build completed successfully
  • The static server is running on port 6006
  • Verify with bun -e "fetch('http://localhost:6006')" (not just curl)

Interactive stories fail with timeout

Radix popovers/tooltips portal to document.body and require precise hover timing. The retry logic handles most cases. If persistent:

  • Run the failing story individually (--story <Name>)
  • Check if the story's mock data or selectors drifted from the actual component

Images look wrong (clipped, misaligned)

  • Check viewport constants in capture-readme-screenshots.ts (1900×1188, DPR 2)
  • Screenshots use viewport-only capture (not fullPage) — if content overflows the viewport, the story layout needs adjustment, not the capture script
  • The StoryDef interface supports a postProcess hook for custom Sharp pipelines if a future story needs cropping/compositing, but no current stories use it

Why there is no ProductHero screenshot story

Intentional — README now uses mux-demo.gif as the hero media. The README screenshot pipeline covers only the 9 docs/img/*.webp assets that remain in README, starting with code-review.webp.

平台分布

Codex

33.98%
按下载量换算83

Claude

30.69%
按下载量换算75

Cursor

16.32%
按下载量换算40

Gemini CLI

9.55%
按下载量换算23

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills