Kenhuang Academy
  • English

      • GitHub Copilot Tutorial

      Welcome to the GitHub Copilot tutorial. Based on the official GitHub Copilot docs, this guide focuses on inline completions, Chat, the coding agent, github.com integration, and Copilot CLI—from first install to a reusable GitHub workflow.

      Copilot is GitHub / Microsoft’s AI coding assistant: it can complete code and chat in your IDE, and it can work Issues and open pull requests on GitHub. UIs, plans, and model lists change quickly—treat the official docs and the Copilot settings on your GitHub account as the source of truth. This course does not invent prices or undocumented shortcuts.

      Table of Contents

      Basics

      1. Introduction — Product fit, plans, vs Cursor and Claude Code
      2. Installation — VS Code extension (optional JetBrains), sign in with GitHub
      3. Quick Start — First ghost-text completion + first Chat

      Intermediate

      1. Completions — Ghost text, comments-to-code, accept / reject
      2. Copilot Chat — Side pane, inline chat, questions about the repo
      3. Coding Agent — Cloud agent: Issue to PR (high-level, official)
      4. GitHub Integration — PRs, Issues, and reviews on github.com
      5. Copilot CLI — Copilot in the terminal

      Practice

      1. Copilot vs Cursor — Honest comparison, and when to use both
      2. Practical Examples — Three tasks, easy to harder
      3. Resources — Official docs and related courses on this site

      Learning Path

      StageGoalChapters
      Day 1Install, sign in, first completion and chatIntro → Install → Quick Start
      Days 2–3Daily edit loopCompletions → Chat
      Week 1Wire Copilot into GitHubAgent → GitHub integration → CLI
      Week 2Choose a stack and practicevs Cursor → Examples → Resources

      Prerequisites

      • Comfort with Git and reading diffs
      • Familiarity with at least one project layout (JavaScript, Python, C#, …)
      • A GitHub account; for teams, confirm Copilot is enabled and that admins have not disabled Chat / Agent
      • Visual Studio Code installed (this course’s main path), or willingness to follow the official JetBrains plugin steps
      • Optional: the intro chapters of this site’s Cursor and Claude Code courses

      How to use this course

      1. Run the official loop first: install the extension → sign in with GitHub → accept one ghost-text suggestion → ask Chat about the current file. That is Installation and Quick Start.
      2. Learn the daily keys: Tab accepts a completion, Esc rejects; Ctrl+Alt+I on Windows / Control+Command+I on macOS opens Chat. Details in Completions and Chat.
      3. Treat GitHub as a second workbench: assign Issues to the cloud agent, ask for PR summaries and reviews. See Coding Agent and GitHub Integration.
      4. Use the CLI for terminal jobs: run copilot at the repo root. See CLI.
      5. Do not fight Cursor over oral rules: stay on Copilot if you are locked into VS Code / GitHub; pick the Cursor course if you want an AI-native editor. Comparison in Copilot vs Cursor.

      Source of truth: docs.github.com/en/copilot. Plan names, allowances, and prices follow Plans for GitHub Copilot and github.com/features/copilot.

      Windows and macOS

      Every chapter notes the two-OS differences (shortcuts, installers, default shells). There is no Linux-only chapter; VS Code, the npm CLI, and github.com on Linux are closer to macOS. If a work laptop is browser-only, GitHub Integration still covers “ask about a file / review a PR”; completions still need an IDE.

      What this course does not cover

      • Exact monthly prices or AI Credit numbers (the official plans page changes)
      • Undocumented model names or private APIs
      • Bypassing org policy, cracked extensions, or putting secrets in prompts
      • Treating a Copilot-opened PR as already human-reviewed

      To build an in-app agent, use this site’s LangChain course; for the protocol, use MCP.

      Next

      评论