Hands-on Labs Overview¶
The labs are designed to be mixed and matched. Start with a fundamentals lab, then move to APIs, tests, agentic workflows, or MCP depending on the audience.
Labs in this workshop¶
| Area | Lab | Language / tools | Typical use |
|---|---|---|---|
| Fundamentals | HTML Image Gallery | HTML, CSS, JavaScript | First Copilot experience, prompt iteration, UI preview. |
| Fundamentals | Rock Paper Scissors | Python | Generate game logic, tests, comments, and simplifications. |
| APIs | Star Wars API in Python | Python, pytest | Data classes, abstract base classes, HTTP calls, tests. |
| APIs | Star Wars API in Java | Java, Maven | DTOs, interfaces, implementations, and unit tests. |
| Agentic / MCP | HTML Image Gallery with MCP | GitHub MCP, Playwright MCP | Create issues, implement from requirements, and test with tools. |
| Agentic / MCP | Build Your Own MCP Server | Python, Java, TypeScript options | Build a weather MCP server and connect it to Copilot. |
| Agentic / MCP | Four in a Row MCP Game | Java, Spring Boot, MCP | Interact with a custom MCP server through Copilot Chat. |
| Extended | Other Labs | Multiple | External labs for additional languages and customer scenarios. |
How to use these labs¶
Each lab includes:
- A scenario and expected outcome.
- Prompts or prompt patterns for Copilot.
- Checkpoints to review generated output.
- Validation steps such as running tests, previewing a page, or inspecting a diff.
Choosing the right lab¶
| Audience | Suggested path |
|---|---|
| New to Copilot | HTML Image Gallery or Rock Paper Scissors. |
| Backend / API developers | Star Wars API in Python or Java. |
| Developers interested in agents | HTML Image Gallery with MCP. |
| Platform or tooling audience | Build Your Own MCP Server. |
| Mixed customer workshop | Start with one fundamentals lab, then offer a language-specific breakout. |
Trainer guidance¶
- Treat generated code as a draft that requires review.
- Prefer small prompt iterations over one large request.
- Explain tool approval before running MCP labs.
- Use Auto model selection where possible and only choose a specific model when a lab needs a capability such as visual input or deeper reasoning.