Skip to content

Latest commit

 

History

History
106 lines (71 loc) · 4.38 KB

part1-setup-and-chat.md

File metadata and controls

106 lines (71 loc) · 4.38 KB

Part 1 – Setup & Exploring Copilot Chat

Navigation: Part 1 | Part 2 – Custom Skills | Part 3 – Custom Agents | Part 4 – Issues, Coding Agent & Review

Setup & Preparation

Goal: Get your environment and repository ready for Copilot experimentation.

1. Find or fork a project

  • Pick a repository you're comfortable working in — any language works. You can:

    • Use an existing project you maintain.
    • Fork a public repo (e.g. from GitHub Explore).

  • Create a new branch, e.g. copilot-hackathon. Keep all experiments in this branch — do not modify your main.

  • Copy the markdown of the relevant part files into issues in your new branch so that you can modify them.

Copilot never commits code without your consent, but keep your hackathon work isolated.

2. Ensure GitHub Copilot is enabled
  • Make sure the GitHub Copilot extension is installed and active in your IDE.
  • Ensure you also have Copilot Chat available (and optionally Copilot Edits/Agent features if supported by your IDE).

Documentation:

You don't need three separate installs — Copilot Chat and Edits are included in most IDE integrations. Just verify both completions and chat are visible in your IDE.

Exploring Copilot Chat Modes

Goal: Learn how Copilot helps you explore, understand, and reason about your code.

3. Explore your code using Ask mode

Pick a file or function you don't know well and experiment freely with Copilot Chat.

Try questions like:

  • "Explain what this function does, step by step."
  • "Where is this class used?"
  • "Could this function be simplified?"
  • "Generate a test for this logic."
  • "Rewrite this using a different algorithm."

Be curious! Ask follow-ups. Ask why. Modify prompts. Try Ask, Explain, Generate, and Edit modes to see how each behaves.

  • Switch between models (e.g. GPT-5 mini, Claude Sonnet 4.6, Gemini 3 Flash) and compare reasoning quality.

Use Copilot Chat to understand code

4. Generate a copilot-instructions.md file

Copilot can analyse your codebase and generate a project instructions file automatically. This file tells Copilot how your project is structured, how to build and run it, and what conventions to follow — so every future chat response is more accurate and consistent.

Why we're doing this

The .github/copilot-instructions.md file acts as a persistent project briefing for Copilot. It summarizes:

  • How the project is structured and what it does
  • How to build, test, and run it
  • Key dependencies, conventions, and architecture decisions

When this file exists, Copilot automatically includes it as context in every request — you don't need to repeat yourself each session.

Steps

Make sure you're on your copilot-hackathon branch before generating the file.

  • Open Copilot Chat in VS Code (Agent mode)
  • Type /init and press Enter
  • Copilot will analyse your workspace and create .github/copilot-instructions.md

Once it's created, open the file and read what Copilot generated. You can edit and expand it freely — any updates will be picked up automatically in future conversations.

Verify the file is loaded as context

  • Open a new Copilot Chat session and ask a question about your project, e.g. "What is this project?"
  • Once Copilot responds, expand the references list at the top of the reply
  • Confirm that .github/copilot-instructions.md is listed — this means it's automatically included as context for every response going forward.

Documentation: Customize AI in VS Code – Set up your project for AI

Next up: Part 2 – Custom Copilot Skills (continues at step 5)