github-copilot-cli

# GitHub Copilot CLI Skill > Skill by [ara.so](https://ara.so) — Devtools Skills collection. GitHub Copilot CLI brings AI-powered development assistance directly to your terminal. It enables code review, debugging, test generation, code explanation, and custom AI agents without leaving the command line. Think of it as having an expert developer available 24/7 in your terminal who can read your codebase, explain patterns, generate code, and help troubleshoot issues. ## Installation ### Prerequisites - GitHub account with Copilot access (free tier, subscription, or education) - GitHub CLI (`gh`) version 2.62.0 or later - Terminal access ### Install GitHub CLI **macOS:** ```bash brew install gh ``` **Windows:** ```bash winget install --id GitHub.cli ``` **Linux (Debian/Ubuntu):** ```bash curl -fsSL https://cli.github.com/packages/githubcli-archive-keyring.gpg | sudo dd of=/usr/share/keyrings/githubcli-archive-keyring.gpg echo "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/githubcli-archive-keyring.gpg] https://cli.github.com/packages stable main" | sudo tee /etc/apt/sources.list.d/github-cli.list > /dev/null sudo apt update sudo apt install gh ``` ### Install Copilot CLI Extension ```bash gh extension install github/gh-copilot ``` ### Authenticate ```bash gh auth login ``` Follow the prompts to authenticate with your GitHub account that has Copilot access. ### Verify Installation ```bash gh copilot --version ``` ## Core Commands ### Interactive Chat Mode Start an interactive conversation with Copilot: ```bash gh copilot ``` This opens a conversational interface where you can ask questions, get code suggestions, and iterate on solutions. ### Quick Explain Get quick explanations without entering interactive mode: ```bash gh copilot explain "git rebase -i HEAD~3" ``` ```bash gh copilot explain "docker-compose up -d" ``` ### Quick Suggest Get command suggestions for tasks: ```bash gh copilot suggest "find all python files modified in last 7 days" ``` ```bash gh copilot suggest "compress all jpg files in current directory" ``` ### Aliases (Shortcuts) Set up convenient aliases: ```bash gh copilot alias -- --shell bash ``` This creates three aliases: - `ghcs` - suggest commands - `ghce` - explain commands - `ghcp` - open Copilot chat After setup, you can use: ```bash ghcs "list all running docker containers" ghce "kubectl get pods -A" ``` ## Interaction Modes ### 1. Chat Mode (Interactive) Best for: Complex problems, iterative development, multi-step workflows ```bash gh copilot ``` Example conversation: ``` > I need to review a Python Flask application for security issues Copilot: I'll help you review your Flask application for security. Could you share: 1. The main application file 2. Any authentication/authorization code 3. Database interaction code > @app.py @auth.py @models.py Copilot: [Analyzes files and provides security review] ``` ### 2. Explain Mode Best for: Understanding unfamiliar commands, code snippets, or error messages ```bash gh copilot explain "awk '{print $2}' file.txt | sort -u" ``` ```bash gh copilot explain "SELECT u.name, COUNT(o.id) FROM users u LEFT JOIN orders o ON u.id = o.user_id GROUP BY u.id" ``` ### 3. Suggest Mode Best for: Getting shell commands for specific tasks ```bash gh copilot suggest "recursively delete all node_modules folders" ``` ```bash gh copilot suggest "monitor CPU and memory usage in real-time" ``` ## Context Awareness ### File Context (@-mentions) Reference specific files in your prompts: ```bash gh copilot > Review @server.py for potential race conditions ``` ```bash gh copilot > Compare @config.dev.json and @config.prod.json and explain differences ``` ### Multi-File Context Reference multiple files for comprehensive analysis: ```bash gh copilot > @models.py @views.py @serializers.py review this Django app for N+1 queries ``` ### Directory Context Analyze entire directories: ```bash gh copilot > @src/ explain the architecture of this application ``` ## Development Workflows ### Code Review ```bash gh copilot > Review @pull_request_changes.diff for: > - Security vulnerabilities > - Performance issues > - Code style violations > - Missing error handling ``` ### Debugging ```bash gh copilot > I'm getting "TypeError: Cannot read property 'map' of undefined" in @app.js line 42 > Here's the error stack: @error.log ``` ### Test Generation For Python: ```bash gh copilot > Generate pytest unit tests for @calculator.py covering edge cases ``` For JavaScript: ```bash gh copilot > Create Jest tests for @api-client.js including error scenarios ``` For Go: ```bash gh copilot > Write table-driven tests for @handler.go ``` ### Refactoring ```bash gh copilot > Refactor @legacy-auth.js to use async/await instead of callbacks ``` ```bash gh copilot > Convert @class-component.jsx to a functional component with hooks ``` ## Custom Agents and Instructions ### Creating Custom Agents Agents are reusable AI assistants with specific expertise. Create an agent configuration: ```bash mkdir -p ~/.config/gh-copilot/agents ``` **Security Review Agent** (`~/.config/gh-copilot/agents/security-reviewer.md`): ```markdown # Security Reviewer Agent You are a security-focused code reviewer specializing in: - OWASP Top 10 vulnerabilities - Input validation and sanitization - Authentication and authorization flaws - SQL injection and XSS prevention - Secrets and credential management Always: 1. Identify specific line numbers with issues 2. Explain the security impact 3. Provide secure code examples 4. Reference relevant OWASP guidelines ``` **Python Expert Agent** (`~/.config/gh-copilot/agents/python-expert.md`): ```markdown # Python Expert Agent You are a Python expert focusing on: - PEP 8 style compliance - Type hints and mypy compatibility - Pythonic idioms and patterns - Performance optimization - Testing with pytest Provide: - Specific improvement suggestions - Code examples following best practices - Performance benchmarks when relevant ``` ### Using Agents ```bash gh copilot > @security-reviewer analyze @app.py ``` ```bash gh copilot > @python-expert review @data_processor.py for optimization opportunities ``` ### Custom Instructions (Global) Set global instructions that apply to all interactions: ```bash mkdir -p ~/.config/gh-copilot ``` Create `~/.config/gh-copilot/instructions.md`: ```markdown # My Development Preferences - **Language**: Prefer TypeScript over JavaScript - **Testing**: Use Jest with React Testing Library - **Style**: Follow Airbnb style guide - **Error Handling**: Always use try-catch and log errors - **Comments**: Add JSDoc comments for public functions - **Security**: Flag any hardcoded credentials or secrets ``` ## Skills (Automated Context) Skills are automatically loaded based on your project structure. ### Creating a Skill Create `.copilot/skills/` directory in your project: ```bash mkdir -p .copilot/skills ``` **Django Project Skill** (`.copilot/skills/django-project.md`): ```markdown # Django Project Configuration ## Stack - Django 4.2 - PostgreSQL 15 - Redis for caching - Celery for async tasks ## Code Standards - Use Django REST Framework serializers - Prefer class-based views - Always use select_related/prefetch_related - Write tests with pytest-django ## Database - Use migrations for all schema changes - Never use raw SQL without parameterization - Connection: via DATABASE_URL environment variable ``` **React Project Skill** (`.copilot/skills/react-conventions.md`): ```markdown # React Project Conventions ## Stack - React 18 with TypeScript - Vite for bundling - React Query for data fetching - Zustand for state management ## Patterns - Functional components with hooks - Custom hooks for reusable logic - Error boundaries for fault tolerance - Suspense for loading states ## File Structure - Components in src/components - Hooks in src/hooks - Utils in src/utils ``` Skills are automatically loaded when you run `gh copilot` in a directory containing `.copilot/skills/`. ## MCP Server Integration Model Context Protocol (MCP) servers extend Copilot CLI with external data sources and APIs. ### Configuration File Create `~/.config/gh-copilot/mcp.json`: ```json { "mcpServers": { "github": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-github"], "env": { "GITHUB_PERSONAL_ACCESS_TOKEN": "${GITHUB_TOKEN}" } }, "postgres": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-postgres"], "env": { "DATABASE_URL": "${DATABASE_URL}" } } } } ``` ### Available MCP Servers **GitHub Integration:** ```json { "github": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-github"], "env": { "GITHUB_PERSONAL_ACCESS_TOKEN": "${GITHUB_TOKEN}" } } } ``` Usage: ```bash gh copilot > Search GitHub for recent issues tagged "bug" in owner/repo ``` **PostgreSQL Database:** ```json { "postgres": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-postgres"], "env": { "DATABASE_URL": "${DATABASE_URL}" } } } ``` Usage: ```bash gh copilot > Query the database for users created in the last 30 days ``` **Filesystem Access:** ```json { "filesystem": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/allowed/directory"] } } ``` **Brave Search:** ```json { "brave-search": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-brave-search"], "env": { "BRAVE_API_KEY": "${BRAVE_API_KEY}" } } } ``` ### Creating Custom MCP Servers MCP servers are command-line tools that implement the MCP protocol. Example structure: ```javascript // my-mcp-server.js import { Server } from '@modelcontextprotocol/sdk/server/index.js'; import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js'; const server = new Server({ name: 'my-custom-server', version: '1.0.0', }); // Define tools server.setRequestHandler('tools/list', async () => ({ tools: [{ name: 'get_data', description: 'Fetch data from custom source', inputSchema: { type: 'object', properties: { query: { type: 'string' } } } }] })); // Implement tool server.setRequestHandler('tools/call', async (request) => { const { name, arguments: args } = request.params; if (name === 'get_data') { const result = await fetchCustomData(args.query); return { content: [{ type: 'text', text: JSON.stringify(result) }] }; } }); const transport = new StdioServerTransport(); await server.connect(transport); ``` Add to `mcp.json`: ```json { "my-custom": { "command": "node", "args": ["/path/to/my-mcp-server.js"] } } ``` ## Real-World Examples ### Example 1: Code Review Workflow ```bash # Start review session gh copilot > I need to review changes before committing. Review @src/auth.py @src/models.py for: > 1. Security issues > 2. Performance problems > 3. Code style > 4. Missing tests # Copilot provides detailed review > Generate pytest tests for the issues you found in @src/auth.py # Copilot generates test code > Now explain how to fix the SQL injection vulnerability you identified