GitHub Copilot can produce a useful function and a convincing mistake with the same confidence. You usually can't tell which one you have until you read the code and run it. A clear request gives the suggestion a target. A real check tells you whether it hit that target.
This module starts with setup in VS Code and a first inline suggestion. From there, you'll choose between Ask, Edit, and Agent according to the work at hand. You'll frame code prompts as testable specifications built from a goal, context, sources, and expectations. Few-shot examples help when a business rule is hard to express in prose. Later lessons apply the same discipline to documentation, tests, and refactoring, then cover privacy safeguards and the route to the GH-300 certification.
You only need VS Code and a project you can safely break. For each change, write down the behavior you expect, read the diff, run the relevant test, and compare the result with the original requirement. Compilation catches syntax and type problems. It can't tell you that valid code does the wrong thing.
What you'll walk away with
Set up GitHub Copilot in VS Code and confirm your first inline suggestion actually works
Choose between Ask, Edit, and Agent modes for the work in front of you
Write code prompts as testable specifications: goal, context, source, and expectations
Use Copilot to document, test, and refactor code without changing its behavior
Separate content exclusions from public-code filtering, and validate every output either way
Map your route to the GH-300 GitHub Copilot certification
