GitHub Copilot Practice Exam

What the GitHub Copilot exam covers

• Using GitHub Copilot Responsibly125 questions
• GitHub Copilot Features233 questions
• Copilot Data and Architecture100 questions
• Prompt Engineering and Context Crafting100 questions
• Developer Productivity with Copilot112 questions
• Privacy, Content Exclusions, and Safeguards124 questions

Free GitHub Copilot sample questions

A sample of 10 questions with answers and explanations. Sign up free to practice all 794.

1. Question 1Using GitHub Copilot ResponsiblySelect all that apply

   An organization drafting a responsible AI policy for its coding assistant wants to cover the right ground. Which three items belong in that policy? Choose three.

   • AA rule capping how many suggestions a developer may accept from the assistant per day
   • BA statement that accountability for merged code remains with the developer and reviewersCorrect
   • CA requirement that reviewers check AI suggested code for bias and inclusive languageCorrect
   • DGuidance on which tasks are appropriate versus inappropriate to delegate to the assistantCorrect
   • EA requirement that the assistant's autocomplete font size match the code editor theme
   ✓ Correct answer: B, C, D

   Substantive responsible AI policy content addresses which tasks suit delegation, requires reviewers to actively check for bias and exclusionary language, and makes clear that people, not the tool, remain accountable for what gets merged. A daily quota on accepted suggestions and a rule about autocomplete font size are administrative or cosmetic details that do not address the ethical risks the policy is meant to manage. Good policy content targets substantive risk rather than arbitrary limits or interface preferences.

   Why the other options are wrong

   • AA daily quota on accepted suggestions is an arbitrary usage limit that does not address the substantive ethical risks a responsible AI policy should cover.
   • EAutocomplete font size is a cosmetic editor preference unrelated to the substantive content a responsible AI policy should address.
2. Question 2Using GitHub Copilot Responsibly

   A company is onboarding new hires who will use GitHub Copilot from day one. Which onboarding practice best supports responsible adoption?

   • AGrant tool access first and cover expectations later informally
   • BAssume prior Copilot experience covers company-specific rules
   • CDirect new hires to figure out the norms from watching teammates
   • DInclude Copilot norms and data rules in onboarding trainingCorrect
   ✓ Correct answer: D

   Embedding Copilot norms, including data handling and review expectations, into structured onboarding ensures consistent understanding from the start. Relying on informal exposure or assumed prior knowledge leaves gaps that only surface after mistakes happen.

   Why the other options are wrong

   • AGranting access before setting expectations invites early habits that are hard to correct later.
   • BPrior experience with the tool elsewhere does not cover this company's specific policies.
   • CLearning purely by observation risks passing on bad habits along with good ones.
3. Question 3GitHub Copilot Features

   A startup with five engineers wants centralized control over who has Copilot access and wants the subscription billed to the company rather than to each engineer personally. Which plan is designed for this scenario?

   • ACopilot Business, purchased and managed by the organizationCorrect
   • BCopilot Pro purchased individually on each engineer's personal account
   • CCopilot Free set up individually on each engineer's personal account
   • DCopilot Pro+ purchased individually on each engineer's personal account
   ✓ Correct answer: A

   Copilot Business is purchased at the organization level, letting an organization owner assign or remove seats for members and consolidating billing under the organization rather than each individual. Free, Pro, and Pro+ are all tied to a single person's personal account and payment method.

   Why the other options are wrong

   • BPro is billed to and managed by the individual subscriber, not an organization.
   • CFree is an individual, unmanaged offering with no centralized seat control.
   • DPro+ is also an individual subscription rather than an organization managed plan.
4. Question 4GitHub Copilot Features

   A technical writer on the team finds Copilot's suggestions distracting while drafting Markdown documentation but wants completions to keep working normally in every other file type. How should this be configured?

   • ASign out of GitHub before opening any Markdown file
   • BDisable the Copilot extension globally for the whole workspace
   • CSet the per language enable setting to false for MarkdownCorrect
   • DUninstall the Copilot extension whenever editing Markdown files
   ✓ Correct answer: C

   Copilot's settings include a per language map where an entry such as markdown can be set to false while a wildcard entry keeps other languages enabled, letting completions stay active everywhere except Markdown. Uninstalling the extension or disabling it globally would remove suggestions everywhere, not just for one file type, and signing out would also block every language, not only Markdown.

   Why the other options are wrong

   • ASigning out removes access to Copilot entirely rather than scoping it to one language.
   • BDisabling the extension globally for the workspace turns off completions everywhere, not only for Markdown.
   • DUninstalling and reinstalling around each Markdown file is impractical and would also remove suggestions for other languages during that time.
5. Question 5GitHub Copilot Features

   A developer is used to Copilot chat suggesting a shell command inside their editor's integrated terminal panel, where they must copy the text and paste it themselves to run it. How does running a command through the standalone Copilot CLI typically differ?

   • AThe standalone CLI can run the approved command itself, without a separate manual paste step.Correct
   • BThe standalone CLI and the editor's integrated terminal panel behave identically in every respect.
   • CThe standalone CLI requires the developer to first save the suggestion to a file before it can be run.
   • DThe standalone CLI only ever displays plain text as a suggestion and cannot run anything on the developer's behalf without it being copied elsewhere.
   ✓ Correct answer: A

   In an editor's chat panel, a suggested terminal command is often just text that the developer copies into a separate terminal to run. The standalone CLI closes that gap by offering to run the command directly once approved, all within the same flow. This is one of the practical differences between the two ways of getting Copilot's help with terminal tasks.

   Why the other options are wrong

   • BThe two experiences are not identical; the manual copy and paste step is exactly what the standalone CLI avoids.
   • CSaving to a file first is not a required step for running an ordinary suggested command through the CLI.
   • DThe standalone CLI does run approved commands directly; it is not limited to only ever displaying inert text.
6. Question 6Copilot Data and Architecture

   A developer working across multiple related files asks why imports and function signatures from neighboring files sometimes appear in the assembled prompt. What purpose do they serve?

   • AThey give the model additional signal about types and APIs in useCorrect
   • BThey replace the need for the current file's own content entirely
   • CThey are required to authenticate the request with the proxy layer
   • DThey are used only to calculate network latency estimates
   ✓ Correct answer: A

   Including imports and function signatures from related files gives the model extra clues about available types, functions, and libraries, which can lead to more accurate completions. This context supplements rather than replaces the current file's content, and it has nothing to do with authentication or latency measurement.

   Why the other options are wrong

   • BNeighboring file content supplements the current file rather than replacing it.
   • CAuthentication with the proxy is handled separately and does not depend on file imports.
   • DLatency estimation is unrelated to why import information is gathered.
7. Question 7Prompt Engineering and Context Crafting

   A security-conscious team wants a way to require that Copilot Chat suggestions in their repository never propose deprecated cryptography functions, and they want this enforced automatically for every contributor, including new hires. What should they configure?

   • AA longer retained chat history so the restriction is eventually inferred over time.
   • BA verbal onboarding note telling new hires to remember to avoid those functions.
   • CA copilot-instructions.md file listing deprecated cryptography functions to avoid.Correct
   • DA branch protection rule that blocks merges without checking actual Copilot usage.
   ✓ Correct answer: C

   Placing the restriction in a .github/copilot-instructions.md file means Copilot Chat automatically incorporates that guidance for anyone working in the repository, without relying on each person remembering it independently. This is far more durable and consistent than informal onboarding notes.

   Why the other options are wrong

   • AChat history is scoped per conversation, so it cannot serve as a durable, team wide standing restriction.
   • BA verbal onboarding note is not enforced by the tool and depends entirely on individual memory.
   • DBranch protection rules govern merge requirements and are unrelated to shaping Copilot suggestions.
8. Question 8Developer Productivity with Copilot

   After Copilot generates a docstring for an existing function, what should the developer do before committing it?

   • ACommit it as-is, since it is inferred from the real signature
   • BDelete the function's original inline comments to avoid duplication
   • CTrust it fully, because a generated docstring cannot be wrong
   • DVerify it matches the real parameters, return, and edge casesCorrect
   ✓ Correct answer: D

   Copilot infers a docstring from the function's signature and body, which usually gives a reasonable first draft. But subtleties like how a function handles empty input or an unusual edge case may not be captured correctly, so the developer should read the generated text against the real logic before committing.

   Why the other options are wrong

   • AInference from the signature is a draft, not a substitute for verifying edge cases.
   • BRemoving existing comments is unrelated to validating the new docstring.
   • CGenerated docstrings can be inaccurate and do need checking.
9. Question 9Privacy, Content Exclusions, and Safeguards

   An enterprise customer asks GitHub's account team what condition must be met for their organization to qualify for GitHub's IP indemnification coverage on Copilot suggested code. What is the correct condition?

   • AThe organization must purchase a separate indemnification add on license
   • BEvery developer must complete a GitHub provided secure coding certification
   • CThe filter that blocks suggestions matching public code must remain enabledCorrect
   • DThe organization must disable Copilot Chat and use only inline completions
   ✓ Correct answer: C

   GitHub extends IP indemnification to Business and Enterprise customers for claims arising from Copilot suggested code, but this coverage depends on the customer keeping the public code matching filter turned on so that suggestions closely matching public code are blocked. Disabling that filter removes the basis for the indemnification for suggestions affected by the change. This ties the legal protection to a concrete, verifiable configuration.

   Why the other options are wrong

   • AThere is no separate paid add on for this coverage; it is included with Business and Enterprise subject to the filter condition.
   • BIndividual developer certification is not a condition of the indemnification; the filter setting is what matters.
   • DDisabling Chat is not required; the condition concerns the public code filter, not which Copilot surface is used.
10. Question 10Privacy, Content Exclusions, and Safeguards

    A developer personally prefers that Copilot never offer completions while editing Markdown files, across every repository they work in, regardless of any admin policy. What is the appropriate way to achieve this?

    • ARename all Markdown files with a different extension so Copilot skips them
    • BAsk an organization owner to add Markdown paths to the org-level exclusion policy
    • CAdjust the per-language enable setting for Copilot in their own editor settingsCorrect
    • DAsk a repository admin to add a Markdown path pattern to content exclusion
    ✓ Correct answer: C

    This is a personal working preference tied to a file type rather than a sensitive-path concern shared across a team, so it belongs in the developer's own editor settings, not in an admin-configured content exclusion rule. Content exclusion is path-based and applies to everyone touching the repository, which is broader than what this developer wants. The per-language toggle exists precisely to let individuals control Copilot's behavior for specific file types without involving repository or organization policy.

    Why the other options are wrong

    • ARenaming files to dodge Copilot is not a supported or reliable way to control suggestions and would break normal tooling.
    • BOrganization-level exclusion is even broader, applying across many repositories and all developers, which overshoots a single person's preference.
    • DRepository-level content exclusion would affect every collaborator in that repo, not just this developer's own sessions.

Related GitHub resources

• GitHub Copilot study guideKey concepts
• GitHub practice examsAll GitHub
• Certification pathWhere this fits
• Certification exam guides & tipsBlog
• Plans & pricingFree & paid
• GitHub Foundations practice examRelated
• GitHub Actions practice examRelated
• GitHub Advanced Security practice examRelated

GitHub Copilot practice exam FAQ