GitHub Copilot Fundamentals
Everything a developer needs before using Copilot seriously: what it is, how to install it, how completions and chat work, what agent mode does, which plan to buy, and how to review what it writes.
The Copilot Stack
Your path from Copilot beginner to agentic software engineer.
Continue learning
path
Every lesson complete.
of lessons complete
Start here
Three questions, answered in your browser. They pick one of the five role-based routes through the lessons that already exist — no new material, no account, nothing sent anywhere.
You chose
Recommended path
Curriculum roadmap
All 8 clusters are published — 100 lessons in all. Each assumes the one before it, so the order is the curriculum rather than a filing system.
Everything a developer needs before using Copilot seriously: what it is, how to install it, how completions and chat work, what agent mode does, which plan to buy, and how to review what it writes.
How Copilot actually behaves in the editor you already use. Feature support differs sharply between VS Code, Visual Studio, JetBrains, Eclipse and Xcode, and this cluster covers each on its own terms — ending with a complete idea-to-pull-request workflow.
Copilot behaves differently in every language, and the difference is not a ranking — it is what the language lets you verify. This cluster covers Python, JavaScript, TypeScript, Java, C#, C++, Go, Rust, PHP, Ruby, Bash and SQL on their own terms: the tooling that catches bad suggestions, the failure modes that recur, and the security review each language actually needs.
Infrastructure is the code where being wrong is most expensive and where the language checks least. This cluster covers Docker, Compose, Kubernetes, Terraform, OpenTofu, Ansible, Linux, AWS, Azure, Google Cloud and GitHub Actions on their own terms — each with a repository-ready example, the local validation sequence that gates it, and an honest record of which commands were actually run.
The terminal agent in depth: sessions, permissions and sandboxing, plan and autopilot modes, custom instructions, custom agents, code review, and controlled CI automation.
The five levers that change Copilot's output: the prompt you write, the context it can see, the instructions that persist, the model behind it, and the validation you apply afterwards. Covers prompt engineering, a hundred reusable prompts, every instruction mechanism and which surfaces support them, prompt files, model selection, and turning team standards into configuration.
An agent is powerful because it can act, which is also why it needs a different trust model. Covers the five agent surfaces and what each can genuinely do, custom agents and the tool lists that constrain them, agent skills, Model Context Protocol end to end, three repository-ready agents you build yourself, and GitHub Agentic Workflows.
Enablement and governance are the same job. Covers Copilot code review and its automation, the security practice AI-assisted code needs, what private repositories actually expose, secret handling, MCP governance at enterprise scale, Business and Enterprise administration, the policy precedence rules that decide which setting wins, and rolling Copilot out so the measurement means something.
Someone who can scope work for an agent, supervise it, review what it produces, and know which parts of the job should not be delegated at all.
Other ways through
The clusters above are ordered by subject, which is right for reference and wrong for “where do I start”. Neither of these adds a lesson — they reorder and test the same ones.
5 routes through the existing lessons, each with its prerequisites, a computed reading time and the labs and projects that go with it.
One per cluster. Ten questions each, scored in your browser, with an explanation for every option and the lessons to revisit for anything you miss. Not a certification.