Don't Panic — GitHub Copilot

GitHub Copilot is a coding assistant that lives where code already lives: in an editor, in chat, and around pull requests. Its original party trick is ghost text. You begin a function, describe an intention in a comment, and a gray suggestion offers to continue. This is not telepathy. It is a very fast prediction based on the context it can see, which is less mystical and considerably easier to correct.

Before this, code completion mostly knew names and syntax. Copilot can propose a function body, explain an error, draft tests, or turn a request into a set of edits. Copilot Chat is for a conversation about code; inline chat keeps that conversation beside the file that needs attention. The useful distinction is scope. A one-line completion is a nudge. A chat request is a proposal. Neither one has quietly taken responsibility for the result, despite appearances to the contrary.

The part that changes the shape of the work is agent mode. It can follow a feature or refactoring across files, suggest commands, and iterate when checks expose a problem. That sounds like a small robot colleague, which is flattering to the robot and unfair to the colleague. The job is still to state the change precisely, inspect the diff, and run the tests that tell you whether the change belongs.

Context is the other load-bearing bit. Custom instructions tell Copilot about conventions and architectural choices. Modular instructions can aim different guidance at different kinds of files. MCP servers can add tools and data sources. More context can make a request more relevant; it can also make a poor rule travel farther. Treat instruction files as team-maintained engineering material, not as decorative wallpaper for an enthusiastic autocomplete.

Start with the tabs that answer the question in front of you. The Introduction explains the interaction modes and their limits. Slides give the map of suggestions, chat, agents, and context. The Cheatsheet holds the shortcuts and configuration locations. Field Notes is for the costs that appear after the demo goes well. The quiz checks whether the distinctions have survived contact with your keyboard.