Token导航 LogoToken导航TokenDH.com
前端设计操作浏览器github未标认证来源可访问许可证需确认审计提醒

create-github-readmecreate GitHub README 前端

Agent Skill

用于围绕 GitHub 仓库、Issue、Pull Request、分支、提交和代码协作流程提供辅助能力。它适合让 Agent 查询项目状态、整理变更、辅助创建或检查协作事项,并把仓库中的信息转成可执行的下一步。使用时需要区分只读查询和写入操作;涉及创建 PR、修改 Issue、推送分支或访问私有仓库时,应确认 token 权限、目标仓库范围和用户授权。

总安装

7,862

周安装

342

GitHub Stars

1

下载量

2,787
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:create-github-readme(create GitHub README 前端)
来源仓库:https://github.com/alfredang/skills
仓库路径:skills/create-github-readme
安装命令:
npx skills add https://github.com/alfredang/skills --skill 'Create GitHub README'
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/alfredang/skills --skill 'Create GitHub README'

简介

create-github-readme 用于生成或更新专业级 GitHub README.md 文件,遵循开放标准最佳实践。

  • 它自动截图并上传项目预览图,若存在旧版 README 则重新生成包含最新截图的版本。
  • 适用于新项目启动或文档维护,提供一键式 README 生成和自动推送功能。
  • 支持多种触发命令(/readme、/create_github_readme 等),适配不同交互习惯。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Create GitHub README

Command

/readme, /create_github_readme, or create_github_readme

Navigate

Documentation

Keywords

readme, github readme, create readme, generate readme, project readme, documentation, readme.md, update readme, write readme, project documentation, create github readme, create_github_readme

Description

Generate or update a professional GitHub README.md file following open standards and best practices. If a README already exists and the user requests an update, regenerates it using the latest template including a fresh screenshot. Auto-pushes to GitHub after generation.

Execution

This skill runs using Claude Code with subscription plan. Do NOT use pay-as-you-go API keys. All AI operations should be executed through the Claude Code CLI environment with an active subscription.

Response

I'll generate (or update) a comprehensive GitHub README.md for the project, then push it to GitHub.

The README will include:

SectionDescription
ScreenshotAuto-captured via Playwright MCP (if live site exists)
HeaderProject name, badges, tagline, links
AboutProject description and key features
Tech StackTechnologies used with badges
ArchitectureSystem diagram (ASCII art)
Project StructureFile/folder organization
Getting StartedPrerequisites, installation, running
DeploymentDocker, Hugging Face Spaces instructions
ContributingFork, PR workflow, discussions
Developed ByTertiary Infotech Academy Pte. Ltd.
AcknowledgementsCredits and thanks

Instructions

Pre-check: Create or Update?

Before generating, check if a README.md already exists:

ls README.md 2>/dev/null

If README.md does NOT exist: Proceed to create a new README from scratch (Section 0 onwards).

If README.md already exists AND the user requests an update:

  1. Read the existing README to understand current content
  2. Check if the README already contains a screenshot (see Section 0.1)
  3. Regenerate the README using the latest template below, preserving project-specific details (description, features, tech stack) from the existing README
  4. Only re-capture the screenshot if the README is missing one (see Section 0)
  5. Overwrite the existing README.md with the updated version

When generating or updating a README, follow this open standard structure:

0. Screenshot Capture (Auto — Conditional)

Before generating the README, determine whether a screenshot capture is needed. A screenshot should only be captured when both conditions are met:

  1. A live demo URL is available
  2. The README does not already contain a screenshot (i.e., no ![Screenshot] or ![screenshot] image reference exists)

If the README already has a screenshot image reference, skip capture entirely and preserve the existing screenshot reference.

0.1 Check for Existing Screenshot in README

If a README.md exists, read it and check for any of these patterns:

  • ![Screenshot](screenshot.png) or similar image markdown with "screenshot" in the alt text or path
  • ![screenshot] image references in any format
  • <img tags referencing a screenshot file

If a screenshot reference is found: Skip to Section 1 — do NOT re-capture. Preserve the existing screenshot section as-is.

If no screenshot reference is found: Continue to Section 0.2 to detect a live URL and capture.

0.2 Detect Live Demo URL

Check these sources in order:

  1. User-provided URL
  2. package.jsonhomepage field
  3. Vercel project URL: https://<project-name>.vercel.app
  4. GitHub Pages URL: https://<owner>.github.io/<repo>/
  5. Any URL found in an existing README.md

0.3 Capture Screenshot with Playwright MCP

Only execute this step if a live demo URL was found in Section 0.2 and no existing screenshot was detected in Section 0.1.

Use the Playwright MCP tools in this sequence:

1. mcp__playwright__browser_navigate → Navigate to the live demo URL
2. mcp__playwright__browser_wait_for → Wait for the page to fully load (wait a few seconds)
3. mcp__playwright__browser_take_screenshot → Take a full-page screenshot and save to project root
4. mcp__playwright__browser_close → Close the browser

Playwright MCP tool calls:

Tool: mcp__playwright__browser_navigate
Args: { "url": "<LIVE_DEMO_URL>" }

Tool: mcp__playwright__browser_wait_for
Args: { "time": 3 }

Tool: mcp__playwright__browser_take_screenshot
Args: { "type": "png", "filename": "screenshot.png", "fullPage": true }

Tool: mcp__playwright__browser_close
Args: {}

Important: The screenshot file will be saved relative to the current working directory. After capture, verify the file exists before referencing it in the README.

0.4 Add Screenshot to README

Insert the screenshot after the header section:

## Screenshot

![Screenshot](screenshot.png)

If the screenshot shows the main UI or hero section, use it. If the page requires authentication or shows a loading state, note this and skip the screenshot.

0.5 If No Live URL Found

Skip screenshot capture and add a commented-out placeholder:

## Screenshot

<!-- Add a screenshot of your app here -->
<!-- ![Screenshot](screenshot.png) -->

1. Header Section

<div align="center">

# Project Name

[![Badge1](url)](link)
[![Badge2](url)](link)

**Tagline**

[Demo](url) · [Report Bug](url) · [Request Feature](url)

</div>

2. Required Badges

  • Python version
  • Framework (Chainlit)
  • AI Provider (Claude/Anthropic)
  • License
  • Docker
  • Deployment platform

3. About Section

  • Brief description (2-3 sentences)
  • Key features as bullet points or table
  • Platform statistics (optional)

4. Tech Stack

Table format with categories:

  • Frontend
  • Backend
  • AI/LLM
  • Database
  • Deployment
  • Document Processing

5. Architecture Diagram

ASCII art showing:

  • UI layer
  • Application layer
  • AI agents
  • Data/document layer

6. Project Structure

project/
├── app.py
├── modules/
└── ...

7. Getting Started

  • Prerequisites
  • Clone & install steps
  • Environment setup
  • Run command

8. Deployment

  • Docker instructions
  • Cloud platform instructions

9. Contributing

  • Fork instructions
  • PR workflow
  • Link to discussions

10. Footer

  • Developed by: Company name
  • Acknowledgements: Credits
  • Call to action: Star the repo

11. Push to GitHub

After generating or updating the README, invoke the /github-push skill to push changes to GitHub.

The github-push skill will:

  1. Scan for exposed secrets (mandatory security check)
  2. Stage README.md and screenshot.png (if created)
  3. Commit with an appropriate message (e.g., docs: generate README or docs: update README to latest template)
  4. Push to the remote repository
  5. Auto-configure repo description, topics, and discussions if needed

Capabilities

  • Auto-capture screenshot of live site using Playwright MCP
  • Generate or update README.md following GitHub best practices
  • Include technology badges with shields.io
  • Create ASCII architecture diagrams
  • Document file structure
  • Provide multi-platform deployment instructions
  • Add proper attribution and acknowledgements
  • Auto-push to GitHub via /github-push skill after generation

Next Steps

After running /create_github_readme:

  1. Verify the push succeeded on GitHub
  2. Review the auto-captured screenshot (screenshot.png) — replace if needed
  3. Verify repo description, topics, and discussions are configured
  4. Share the repository link

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

34.71%
按下载量换算967

Claude

31.45%
按下载量换算877

Cursor

16.86%
按下载量换算470

Gemini CLI

8.48%
按下载量换算236

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills