This repository shows how GitHub Copilot helps you move quickly when you’re building something new — the kind of fast, iterative work you’d do during a hackathon. It uses one small project-submission service as a throughline, so you can see Copilot support the whole flow rather than a set of disconnected tricks.
Follow along in the companion repository: katiem0/copilot-build-faster-demo.
What You’ll Learn
- How to use Copilot to scaffold new features quickly when you’re short on time
- How Copilot helps you debug — both from a failing test and from unexpected runtime behavior
- How to generate meaningful tests and expand realistic sample data
- How to run an in-editor code review and make it repeatable with custom prompts
- Habits that keep Copilot’s output safe and consistent
Prerequisites
- An active GitHub Copilot subscription
- VS Code with the GitHub Copilot and Copilot Chat extensions
- Node.js 18+ to run the sample service (Express, with Jest and Supertest for tests)
- The
copilot-build-faster-demorepository cloned and set up (npm install)
What the Demos Show
The repository walks through the workflows you actually hit when building under pressure, each demonstrated against the same service:
- Scaffolding fast — adding a small feature and a focused test quickly, so you can see Copilot accelerate the “get something working” phase.
- Debugging with a test — writing a regression test that exposes a real gap in the code, then letting Copilot trace the failure to its root cause.
- Debugging without a test — exercising the running API by hand, capturing an unexpected response, and asking Copilot to reason through the route, store, and validation flow.
- Making sense of tooling errors — interpreting terminal, dependency, or compiler-style failures and applying the smallest fix.
- Tests and sample data — generating edge-case tests and extending synthetic fixtures so you have realistic traffic to demo with.
- In-editor code review — reviewing a change with Copilot and comparing the findings against a reusable review prompt.
The project includes an intentional gap — duplicate team members aren’t rejected — so there’s a real bug for Copilot to find and fix during the debugging demos.
Getting the Most From It
- Read the repo’s example instructions and prompt files to see how a little up-front setup keeps Copilot’s output consistent and cuts down on re-prompting.
- Use synthetic data only — never paste real names, secrets, tokens, or customer data into prompts.
- Review generated code before accepting it, and run the tests after each Copilot-assisted change.
Key Takeaway
Copilot shines when you’re moving fast — it helps you scaffold, debug, test, and review without breaking your flow, as long as you stay in the driver’s seat and verify what it produces.