Keeping the Workshop Current¶
GitHub Copilot changes too quickly for a static workshop to maintain a reliable product roadmap. Instead of predicting future features, this page gives trainers and maintainers a repeatable way to keep the workshop aligned with the current Copilot experience.
What to review before each delivery¶
| Area | What to check |
|---|---|
| Copilot UI | Mode names, chat controls, tool approval prompts, and screenshot accuracy. |
| Models | Current model names, Auto behavior, visual-input support, and organization policy. |
| MCP | VS Code MCP configuration, server startup UX, and tool enablement screens. |
| Lab code | Dependencies, APIs, build commands, tests, and runtime versions. |
| External labs | Repository availability, instructions, QR codes, and prerequisites. |
| Security guidance | Token handling, generated-code review, dependency review, and customer data boundaries. |
Current capability areas to watch¶
- Agentic coding workflows: multi-file changes, terminal commands, tests, and tool use with review.
- MCP servers: bringing external tools and data sources into Copilot Chat.
- Model selection: Auto model selection, specialized reasoning models, and visual input support.
- Enterprise controls: policy, auditability, content exclusion, and organization-level model availability.
- Repository context: how Copilot uses open files, selections, workspace search, issues, pull requests, and docs.
Recommended maintenance workflow¶
- Review the official Copilot docs and changelog.
- Run the selected labs from a clean clone.
- Update prompts and screenshots when the product UX changes.
- Replace hard-coded model names with task-based guidance unless a lab requires a specific model.
- Build the MkDocs site and fix broken links or assets.
- Commit changes in the fork and open a pull request to the Azure-Samples repository.
Official sources¶
- GitHub Copilot documentation
- AI model comparison for GitHub Copilot
- GitHub Changelog
- VS Code AI documentation
- Model Context Protocol documentation
Guidance for trainers¶
When a participant sees a slightly different UI, treat it as normal. Copilot capabilities can vary by plan, organization policy, extension version, IDE, and rollout status. Focus on durable practices: provide context, ask for small iterations, inspect diffs, run tests, and keep humans in control of accepted changes.