GH-300 Official Certification Question Bank (125 Questions)
125 curated questions sourced directly from official Microsoft Learn study domains, GH-300 exam prep guides, and real test sets covering all core competency areas of the GH-300 exam.
Q71: On Copilot Individual, how are user prompts used if the telemetry sharing setting is left ON?
If left ON, GitHub uses prompt and completion telemetry to analyze usage patterns, evaluate system performance, and fine-tune base AI models. Users can opt out in profile settings.
- Opt-in data helps fine-tune general models.
- Can be disabled at any time in individual profile settings.
What is telemetry data used for when enabled on a Copilot Individual account?
Q72: Can Content Exclusion rules block Copilot from working on specific IP addresses or networks?
No. Content Exclusion rules target repository file paths and URLs. Network-level access is managed via enterprise SAML SSO and IP allow lists.
- Content Exclusions target file paths and repository URLs.
- Network access is controlled via SAML/IP allow lists.
What resource type is targeted by Content Exclusion rules in GitHub Copilot?
Q73: When should a developer use Inline Chat over the Sidebar Chat panel?
Use Inline Chat (Cmd/Ctrl + I) for targeted localized refactoring or code generation on specific lines. Use Sidebar Chat for broad architectural queries, @workspace searches, or multi-file context requests.
- Inline Chat: Localized, fast edits directly on selected lines.
- Sidebar Chat: Broad queries, multi-file context, and workspace searches.
Which interface is best suited for localized inline refactoring on 5 highlighted lines?
Q74: What is 'Role Prompting' and how does it affect Copilot Chat responses?
Role Prompting assigns a persona or specialized role to the AI (e.g., 'Act as a Senior Cyber Security Auditor'), focusing the response on role-specific standards and security checks.
- Role Prompting frames responses using target expertise personas.
- Helps guide tone, focus, and domain-specific depth.
Phrasing a prompt as 'Act as a Senior Database Administrator' is an example of what technique?
Q75: What integrated mechanisms in GitHub Enterprise work alongside Copilot to detect vulnerabilities introduced into repositories?
GitHub Copilot pairs with GitHub Advanced Security (GHAS) features—such as Code Scanning (CodeQL), Secret Scanning, and Dependabot—to catch vulnerabilities or leaked secrets before code reaches production.
- Copilot generates code proposals; GHAS validates security.
- CodeQL catches SAST security flaws; Secret Scanning blocks leaked API keys.
Which GitHub feature family provides automated static application security testing (SAST) for Copilot-generated code?
Q76: What underlying technology powers GitHub Copilot's code generation capabilities?
GitHub Copilot is powered by advanced Large Language Models (LLMs) customized for software development (including OpenAI GPT models tuned on code and fine-tuned by GitHub).
- Backed by OpenAI Large Language Models fine-tuned on code.
- Processes multi-lingual programming and natural language prompts.
What class of AI models powers GitHub Copilot?
Q77: How can developers see which specific files Copilot Chat used to construct its response?
Copilot Chat displays a 'Used X references' collapsible accordion at the top of its response, listing the exact file paths and line ranges evaluated as context.
- References list shows exact context files evaluated.
- Allows developers to verify source attribution and context accuracy.
Where can you inspect the list of project files Copilot used to generate a Chat response?
Q78: What is the formal definition of an 'AI Hallucination' in generative tools?
An AI Hallucination refers to outputs generated by an LLM that are syntactically plausible but factually incorrect, logically invalid, or referencing non-existent real-world entities.
- Plausible-sounding but factually wrong or non-existent outputs.
- Requires human verification across generative outputs.
Which phrase describes an AI hallucination?
Q79: How do you trigger terminal inline command generation in VS Code?
Focusing the integrated terminal and pressing Cmd + I (macOS) or Ctrl + I (Windows/Linux) opens inline chat inside the terminal pane to generate shell commands.
Cmd/Ctrl + Iinside terminal launches terminal inline chat.- Generates bash, zsh, or PowerShell command strings directly.
What shortcut opens inline chat inside the integrated terminal pane?
Q80: What is 'Context Stuffing' and why does it negatively impact Copilot performance?
Context Stuffing refers to dumping large volumes of un-related source files into a prompt. This clutters the token budget and lowers completion accuracy by introducing irrelevant noise.
- Avoid adding unrelated project files to prompts.
- Focus prompts strictly on relevant files and requirements.
Why should developers avoid including unrelated files when prompting Copilot?