Instructor: Tim Warner | GH-300 | Four-hour O'Reilly Live Learning course
Thursday, September 24, 2026 | 9 a.m.-1 p.m. Central / 7-11 a.m. Pacific
Welcome! This is the public learner repository for the September 24, 2026 class and for independent study afterward. Our route is simple: validate AI output, choose the right Copilot surface, improve the context, and apply the right safeguards.
Exam baseline: skills measured August 7, 2026. Product behavior and links reviewed September 20, 2026. The August update is a minor refinement of the six-domain blueprint, not a new exam.
- Read the course plan and quick reference.
- Work through the class activities. Each has an observable outcome and a no-account route.
- Use Cert Buddy for original practice, finish the mini mock, then work the 60-item practice bank.
- Follow the study guide, the objective-mapped resource links, and official exam links for your remaining study.
| Segment | Central time | What you will do |
|---|---|---|
| 1. Foundations and responsible operation | 9-10 a.m. | Review flawed generated code and decide what evidence makes a change trustworthy |
| 2. Core features: IDE, Chat, CLI | 10-11 a.m. | Compare two Copilot surfaces and reuse a prompt for two tasks |
| 3. Data, agents, MCP, and context | 11 a.m.-noon | Improve a prompt, retrieve documentation, and explain session/context choices |
| 4. Privacy, governance, productivity, and exam practice | Noon-1 p.m. | Match safeguards to risks, interpret metrics, take the mini mock, and plan further study |
The first three segments include a five-minute break. The last five minutes of Segment 4 are the wrap-up and tonight's study plan. See the course plan for demo entry points and fallbacks.
| Resource | Purpose |
|---|---|
| Current course plan | Four teaching segments, demo entry points, timings, and rehearsal gates |
| Class activities | Responsible AI, Chat/CLI, context, testing, and policy practice |
| GH-300 objectives | Six domains and the current skills measured |
| Study guide | What to understand and what evidence to produce |
| Resource links | Every objective mapped to verified Microsoft Learn and GitHub Docs sources |
| Practice bank | 60 original items weighted to the blueprint, with per-choice rationale |
| Hooks and governance lab | A tool-use logger, approval boundaries, and enterprise policy walkthrough |
| Metrics tour | Current report format; synthetic demo needs no credentials |
| Tips app | Small Node.js application to inspect, test, and improve |
| September teaching deck | Updated class slides and presenter notes |
| Refresh notes | Changes, sources, and what still needs live rehearsal |
Recommended: current VS Code with GitHub Copilot access, Git, and Node.js 22 or later. Windows examples use PowerShell 7. Copilot availability depends on your account and organization policy. An enterprise license is not required for the core activities.
Before trusting a downloaded workspace, inspect its instructions, MCP configuration, and executable hooks. This repository includes teaching hooks in .github/hooks/; their scripts are readable in scripts/hooks/.
# Use your own copy so classroom edits remain separate from the published materials.
git clone https://github.com/timothywarner-org/copilot-cert-prep.git
Set-Location copilot-cert-prep
code .The tips app and synthetic metrics demo use Node.js built-ins and need no package installation:
node src/app.js
node copilot-metrics-tour/index.js --demoFor the repository's automated checks:
npm ci
npm test -- --runInBand
npm run check:contentOptional Copilot CLI setup, from the official installation guide:
# Node.js 22+ is required by this installation method.
npm install -g @github/copilot
copilot
# Inside Copilot CLI, enter /login if authentication is required.If you cannot install software or use Copilot, read the supplied scenarios, write your decisions, and compare with the debriefs. You can meet the learning outcomes without changing a tenant.
Open Copilot Chat and select gh300-cert-buddy-agent from the agent picker. Its definition is in the agent file; its three skills cover questions, labs, and study plans.
The Microsoft Learn MCP server is preconfigured in .vscode/mcp.json as gh300buddy-mslearn. It uses the public Learn endpoint without an API key. Start/authorize it through VS Code's MCP controls after inspecting the configuration. The separate GitHub MCP entry is optional and requires your own authentication; Cert Buddy does not need it for Learn retrieval.
Try:
Give me one original GH-300 question about content exclusions. Cite current primary documentation after I answer.
Build a 10-minute Copilot CLI exercise with an explicit success check and cleanup.
Help me plan tonight's study. I am weak on data flow and confident with inline suggestions.
Cert Buddy must show the question first, wait for your answer, and then explain all four choices with sources. Its validator checks structure and flags length giveaways; it cannot prove factual accuracy. If retrieval is unavailable, the agent must disclose that limitation.
| GH-300 domain | Weight |
|---|---|
| Use GitHub Copilot responsibly | 15-20% |
| Use GitHub Copilot features | 25-30% |
| Understand GitHub Copilot data and architecture | 10-15% |
| Apply prompt engineering and context crafting | 10-15% |
| Improve developer productivity with GitHub Copilot | 10-15% |
| Configure privacy, content exclusions, and safeguards | 10-15% |
Study decisions, not a model roster. Model availability, billing, previews, and UI details change faster than the exam objectives. Hooks and server-managed settings are useful governance enrichment; the blueprint does not name them as separate objectives.
Use the linked current Markdown sources and September deck for this class. June-named Markdown pages redirect older bookmarks to their current equivalents; superseded materials remain available in Git history. This repository contains original teaching material, not real exam questions or a guarantee of passing.
Code and course materials in this repository are released under the MIT License. The O'Reilly name and logo, the GitHub Copilot product interface shown in screenshots, and the GH-300 objectives quoted from Microsoft Learn belong to their respective owners and are used here for instruction.