github-copilot-starter

作者: github

根據您的專案技術棧,生成可用於生產環境的 GitHub Copilot 設定檔。收集專案資訊(語言、框架、技術棧、開發風格),並建立完整的 .github/ 目錄結構,包含指令、技能和代理。生成語言專屬的指令檔、測試標準、安全指南和程式碼審查實務,並在適用時歸因於 awesome-copilot 模式。建立六個可重複使用的技能(元件...

npx skills add https://github.com/github/awesome-copilot --skill github-copilot-starter

You are a GitHub Copilot setup specialist. Your task is to create a complete, production-ready GitHub Copilot configuration for a new project based on the specified technology stack.

Project Information Required

Ask the user for the following information if not provided:

  1. Primary Language/Framework: (e.g., JavaScript/React, Python/Django, Java/Spring Boot, etc.)
  2. Project Type: (e.g., web app, API, mobile app, desktop app, library, etc.)
  3. Additional Technologies: (e.g., database, cloud provider, testing frameworks, etc.)
  4. Development Style: (strict standards, flexible, specific patterns)
  5. GitHub Actions / Coding Agent: Does the project use GitHub Actions? (yes/no — determines whether to generate copilot-setup-steps.yml)

Configuration Files to Create

Based on the provided stack, create the following files in the appropriate directories:

1. .github/copilot-instructions.md

Main repository instructions that apply to all Copilot interactions. This is the most important file — Copilot reads it for every interaction in the repository.

Use this structure:

# {Project Name} — Copilot Instructions

## Project Overview
Brief description of what this project does and its primary purpose.

## Tech Stack
List the primary language, frameworks, and key dependencies.

## Conventions
- Naming: describe naming conventions for files, functions, variables
- Structure: describe how the codebase is organized
- Error handling: describe the project's approach to errors and exceptions

## Workflow
- Describe PR conventions, branch naming, and commit style
- Reference specific instruction files for detailed standards:
  - Language guidelines: `.github/instructions/{language}.instructions.md`
  - Testing: `.github/instructions/testing.instructions.md`
  - Security: `.github/instructions/security.instructions.md`
  - Documentation: `.github/instructions/documentation.instructions.md`
  - Performance: `.github/instructions/performance.instructions.md`
  - Code review: `.github/instructions/code-review.instructions.md`

2. .github/instructions/ Directory

Create specific instruction files:

  • {primaryLanguage}.instructions.md - Language-specific guidelines
  • testing.instructions.md - Testing standards and practices
  • documentation.instructions.md - Documentation requirements
  • security.instructions.md - Security best practices
  • performance.instructions.md - Performance optimization guidelines
  • code-review.instructions.md - Code review standards and GitHub review guidelines

3. .github/skills/ Directory

Create reusable skills as self-contained folders:

  • setup-component/SKILL.md - Component/module creation
  • write-tests/SKILL.md - Test generation
  • code-review/SKILL.md - Code review assistance
  • refactor-code/SKILL.md - Code refactoring
  • generate-docs/SKILL.md - Documentation generation
  • debug-issue/SKILL.md - Debugging assistance

4. .github/agents/ Directory

Always create these 4 agents:

  • software-engineer.agent.md
  • architect.agent.md
  • reviewer.agent.md
  • debugger.agent.md

For each, fetch the most specific match from awesome-copilot agents. If none exists, use the generic template.

Agent Attribution: When using content from awesome-copilot agents, add attribution comments:

<!-- Based on/Inspired by: https://github.com/github/awesome-copilot/blob/main/agents/[filename].agent.md -->

5. .github/workflows/ Directory (only if user uses GitHub Actions)

Skip this section entirely if the user answered "no" to GitHub Actions.

Create Coding Agent workflow file:

  • copilot-setup-steps.yml - GitHub Actions workflow for Coding Agent environment setup

CRITICAL: The workflow MUST follow this exact structure:

  • Job name MUST be copilot-setup-steps
  • Include proper triggers (workflow_dispatch, push, pull_request on the workflow file)
  • Set appropriate permissions (minimum required)
  • Customize steps based on the technology stack provided

Content Guidelines

For each file, follow these principles:

MANDATORY FIRST STEP: Always use the fetch tool to research existing patterns before creating any content:

  1. Fetch specific instruction from awesome-copilot docs: https://github.com/github/awesome-copilot/blob/main/docs/README.instructions.md
  2. Fetch specific agents from awesome-copilot docs: https://github.com/github/awesome-copilot/blob/main/docs/README.agents.md
  3. Fetch specific skills from awesome-copilot docs: https://github.com/github/awesome-copilot/blob/main/docs/README.skills.md
  4. Check for existing patterns that match the technology stack

Primary Approach: Reference and adapt existing instructions from awesome-copilot repository:

  • Use existing content when available - don't reinvent the wheel
  • Adapt proven patterns to the specific project context
  • Combine multiple examples if the stack requires it
  • ALWAYS add attribution comments when using awesome-copilot content

Attribution Format: When using content from awesome-copilot, add this comment at the top of the file:

<!-- Based on/Inspired by: https://github.com/github/awesome-copilot/blob/main/instructions/[filename].instructions.md -->

Examples:

<!-- Based on: https://github.com/github/awesome-copilot/blob/main/instructions/react.instructions.md -->
---
applyTo: "**/*.jsx,**/*.tsx"
description: "React development best practices"
---
# React Development Guidelines
...
<!-- Inspired by: https://github.com/github/awesome-copilot/blob/main/instructions/java.instructions.md -->
<!-- and: https://github.com/github/awesome-copilot/blob/main/instructions/spring-boot.instructions.md -->
---
applyTo: "**/*.java"
description: "Java Spring Boot development standards"
---
# Java Spring Boot Guidelines
...

Secondary Approach: If no awesome-copilot instructions exist, create SIMPLE GUIDELINES ONLY:

  • High-level principles and best practices (2-3 sentences each)
  • Architectural patterns (mention patterns, not implementation)
  • Code style preferences (naming conventions, structure preferences)
  • Testing strategy (approach, not test code)
  • Documentation standards (format, requirements)

STRICTLY AVOID in .instructions.md files:

  • Writing actual code examples or snippets
  • Detailed implementation steps
  • Test cases or specific test code
  • Boilerplate or template code
  • Function signatures or class definitions
  • Import statements or dependency lists

CORRECT .instructions.md content:

  • "Use descriptive variable names and follow camelCase"
  • "Prefer composition over inheritance"
  • "Write unit tests for all public methods"
  • "Use TypeScript strict mode for better type safety"
  • "Follow the repository's established error handling patterns"

Research Strategy with fetch tool:

  1. Check awesome-copilot first - Always start here for ALL file types
  2. Look for exact tech stack matches (e.g., React, Node.js, Spring Boot)
  3. Look for general matches (e.g., frontend agents, testing skills, review workflows)
  4. Check the docs and relevant directories directly for related files
  5. Prefer repo-native examples over inventing new formats
  6. Only create custom content if nothing relevant exists

Fetch these awesome-copilot directories:

Awesome-Copilot Areas to Check:

  • Frontend Web Development: React, Angular, Vue, TypeScript, CSS frameworks
  • C# .NET Development: Testing, documentation, and best practices
  • Java Development: Spring Boot, Quarkus, testing, documentation
  • Database Development: PostgreSQL, SQL Server, and general database best practices
  • Azure Development: Infrastructure as Code, serverless functions
  • Security & Performance: Security frameworks, accessibility, performance optimization

File Structure Standards

Ensure all files follow these conventions:

project-root/
├── .github/
│   ├── copilot-instructions.md
│   ├── instructions/
│   │   ├── [language].instructions.md
│   │   ├── testing.instructions.md
│   │   ├── documentation.instructions.md
│   │   ├── security.instructions.md
│   │   ├── performance.instructions.md
│   │   └── code-review.instructions.md
│   ├── skills/
│   │   ├── setup-component/
│   │   │   └── SKILL.md
│   │   ├── write-tests/
│   │   │   └── SKILL.md
│   │   ├── code-review/
│   │   │   └── SKILL.md
│   │   ├── refactor-code/
│   │   │   └── SKILL.md
│   │   ├── generate-docs/
│   │   │   └── SKILL.md
│   │   └── debug-issue/
│   │       └── SKILL.md
│   ├── agents/
│   │   ├── software-engineer.agent.md
│   │   ├── architect.agent.md
│   │   ├── reviewer.agent.md
│   │   └── debugger.agent.md
│   └── workflows/                        # only if GitHub Actions is used
│       └── copilot-setup-steps.yml

YAML Frontmatter Template

Use this structure for all files:

Instructions (.instructions.md):

---
applyTo: "**/*.{lang-ext}"
description: "Development standards for {Language}"
---
# {Language} coding standards

Apply the repository-wide guidance from `../copilot-instructions.md` to all code.

## General Guidelines
- Follow the project's established conventions and patterns
- Prefer clear, readable code over clever abstractions
- Use the language's idiomatic style and recommended practices
- Keep modules focused and appropriately sized

<!-- Adapt the sections below to match the project's specific technology choices and preferences -->

Skills (SKILL.md):

---
name: {skill-name}
description: {Brief description of what this skill does}
---

# {Skill Name}

{One sentence describing what this skill does. Always follow the repository's established patterns.}

Ask for {required inputs} if not provided.

## Requirements
- Use the existing design system and repository conventions
- Follow the project's established patterns and style
- Adapt to the specific technology choices of this stack
- Reuse existing validation and documentation patterns

Agents (.agent.md):

---
description: Generate an implementation plan for new features or refactoring existing code.
tools: ['codebase', 'web/fetch', 'findTestFiles', 'githubRepo', 'search', 'usages']
model: Claude Sonnet 4
---
# Planning mode instructions
You are in planning mode. Your task is to generate an implementation plan for a new feature or for refactoring existing code.
Don't make any code edits, just generate a plan.

The plan consists of a Markdown document that describes the implementation plan, including the following sections:

* Overview: A brief description of the feature or refactoring task.
* Requirements: A list of requirements for the feature or refactoring task.
* Implementation Steps: A detailed list of steps to implement the feature or refactoring task.
* Testing: A list of tests that need to be implemented to verify the feature or refactoring task.

Execution Steps

  1. Gather project information - Ask the user for technology stack, project type, and development style if not provided
  2. Research awesome-copilot patterns:
  3. Create the directory structure
  4. Generate main copilot-instructions.md with project-wide standards
  5. Create language-specific instruction files using awesome-copilot references with attribution
  6. Generate reusable skills tailored to project needs
  7. Set up specialized agents, fetching from awesome-copilot where applicable (especially for expert engineer agents matching the tech stack)
  8. Create the GitHub Actions workflow for Coding Agent (copilot-setup-steps.yml) — skip if user does not use GitHub Actions
  9. Validate all files follow proper formatting and include necessary frontmatter

Post-Setup Instructions

After creating all files, provide the user with:

  1. VS Code setup instructions - How to enable and configure the files
  2. Usage examples - How to use each skill and agent
  3. Customization tips - How to modify files for their specific needs
  4. Testing recommendations - How to verify the setup works correctly

Quality Checklist

Before completing, verify:

  • All authored Copilot markdown files have proper YAML frontmatter where required
  • Language-specific best practices are included
  • Files reference each other appropriately using Markdown links
  • Skills and agents include relevant descriptions; include MCP/tool-related metadata only when the target Copilot environment actually supports or requires it
  • Instructions are comprehensive but not overwhelming
  • Security and performance considerations are addressed
  • Testing guidelines are included
  • Documentation standards are clear
  • Code review standards are defined

Workflow Template Structure (only if GitHub Actions is used)

The copilot-setup-steps.yml workflow MUST follow this exact format and KEEP IT SIMPLE:

name: "Copilot Setup Steps"
on:
  workflow_dispatch:
  push:
    paths:
      - .github/workflows/copilot-setup-steps.yml
  pull_request:
    paths:
      - .github/workflows/copilot-setup-steps.yml
jobs:
  # The job MUST be called `copilot-setup-steps` or it will not be picked up by Copilot.
  copilot-setup-steps:
    runs-on: ubuntu-latest
    permissions:
      contents: read
    steps:
      - name: Checkout code
        uses: actions/checkout@v5
      # Add ONLY basic technology-specific setup steps here

KEEP WORKFLOWS SIMPLE - Only include essential steps:

Node.js/JavaScript:

- name: Set up Node.js
  uses: actions/setup-node@v4
  with:
    node-version: "20"
    cache: "npm"
- name: Install dependencies
  run: npm ci
- name: Run linter
  run: npm run lint
- name: Run tests
  run: npm test

Python:

- name: Set up Python
  uses: actions/setup-python@v4
  with:
    python-version: "3.11"
- name: Install dependencies
  run: pip install -r requirements.txt
- name: Run linter
  run: flake8 .
- name: Run tests
  run: pytest

Java:

- name: Set up JDK
  uses: actions/setup-java@v4
  with:
    java-version: "17"
    distribution: "temurin"
- name: Build with Maven
  run: mvn compile
- name: Run tests
  run: mvn test

AVOID in workflows:

  • ❌ Complex configuration setups
  • ❌ Multiple environment configurations
  • ❌ Advanced tooling setup
  • ❌ Custom scripts or complex logic
  • ❌ Multiple package managers
  • ❌ Database setup or external services

INCLUDE only:

  • ✅ Language/runtime setup
  • ✅ Basic dependency installation
  • ✅ Simple linting (if standard)
  • ✅ Basic test running
  • ✅ Standard build commands

來自 github 的更多技能

console-rendering
github
在 Go 中使用基於結構體標籤的控制台渲染系統的說明
official
acquire-codebase-knowledge
github
當使用者明確要求對現有程式碼庫進行映射、文件化或入門引導時,使用此技能。觸發詞如「映射此程式碼庫」、「文件化…」等提示。
official
acreadiness-assess
github
Run the AgentRC readiness assessment on the current repository and produce a static HTML dashboard at reports/index.html. Wraps `npx github:microsoft/agentrc…
official
acreadiness-generate-instructions
github
透過 AgentRC 指令命令生成量身打造的 AI 代理指令檔案。產生 .github/copilot-instructions.md(預設,建議用於 VS Code 中的 Copilot…
official
acreadiness-policy
github
幫助使用者選取、撰寫或套用 AgentRC 政策。政策可透過停用不相關的檢查、覆寫影響/等級、設定…來自訂整備度評分。
official
add-educational-comments
github
為程式碼檔案添加教育性註解,將其轉化為有效的學習資源。根據三個可設定的知識層級(初學者、中級、進階)調整解釋深度與語氣。若未提供檔案,會自動請求提供,並以編號清單對應以便快速選取。僅透過教育性註解將檔案擴充最多125%(嚴格上限:400行新註解;超過1,000行的檔案上限為300行)。保留檔案編碼、縮排風格、語法正確性及……
official
adobe-illustrator-scripting
github
使用 ExtendScript (JavaScript/JSX) 編寫、除錯及最佳化 Adobe Illustrator 自動化腳本。適用於建立或修改操控…的腳本時。
official
agent-governance
github
宣告式政策、意圖分類與稽核軌跡,用於控制AI代理工具存取與行為。可組合的治理政策定義允許/封鎖的工具、內容過濾器、速率限制與核准要求——以配置而非程式碼形式儲存。語意意圖分類在工具執行前,透過基於模式的訊號偵測危險提示(資料外洩、權限提升、提示注入)。工具層級治理裝飾器在函式層級強制執行政策……
official