Microsoft GH-600 : Developing in Agentic AI Systems

You have a GitHub Actions workflow that starts multiple agent jobs in parallel.
You need to prevent two runs in the same job from executing simultaneously on the same branch. The solution must ensure, at most, one running instance and one pending instance per branch.
What should you add to the job?

• A. jobs.<job_id>.concurrency.cancel-in-progress: ${{ !contains(github.ref, 'release/') }}
• B. jobs.<job_id>.concurrency.group: ci-${{ github.ref }}
• C. jobs.<job_id>.strategy.max-parallel: 1
• D. jobs.<job_id>.strategy.matrix: [1]

You have a GitHub repository that uses three GitHub Copilot coding agents named agent1, agent2, and agent3.
During structured evaluation runs, agent1 frequently returns Markdown narratives instead of a machine-parsable result.
You need to ensure that agent1 consistently returns only a predefined JSON structure without affecting the output of the other agents.
Which file should you modify?

• A. .github/instructions/agent1.instructions.md
• B. .github/copilot-instructions.md
• C. .github/agents/agent1.agent.md
• D. .github/workflows/copilot-setup-steps.yml

You have a GitHub Copilot coding agent named CodeAgent. The .agent.md file of CodeAgent contains the following YAML frontmatter.
name: CodeAgent
description: Performs repository analysis and code review tasks.
tools: ['edit', 'execute', 'read', 'search']
You need to issue a GitHub Copilot CLI command that preserves execution velocity for read-only tasks by eliminating approval prompts for low-risk tools. The solution must ensure that high-risk tools that can make changes remain available but still require explicit user approval before running.
Which command should you run?

• A. copilot agent run CodeAgent
• B. copilot agent run CodeAgent --deny-tool 'edit,execute'
• C. copilot agent run CodeAgent --allow-tool 'read,search'
• D. copilot agent run CodeAgent --allow-all-tools

You have a GitHub repository that uses GitHub Actions to run an autonomous agent job. The job summarizes a pull request and posts the summary as a pull request comment.
Your company stores required configuration values, such as JOB_NAME and RUNNER, as organization-level configuration variables and exposes them in workflows by using the vars context.
You discover that the agent job runs on push events and creates comments that are NOT tied to a pull request.
You need to ensure that the agent job runs only when the workflow run is triggered by a pull request. The solution must keep the job name and runner selection configured by using variables.
What should you do in the workflow?

• A. Set jobs.agent.if: ${{ env.GITHUB_EVENT_NAME == 'pull_request' }}. Set jobs.agent.name and jobs.agent.runs-on by using the env context.
• B. Set jobs.agent.if: ${{ github.event_name == 'push' }}. Set jobs.agent.name and jobs.agent.runs-on by using the vars context.
• C. Set jobs.agent.if: ${{ github.event_name == 'pull_request' }}. Set jobs.agent.name and jobs.agent.runs-on by using the vars context.
• D. Set jobs.agent.if: ${{ github.ref == 'refs/heads/main' }}. Set jobs.agent.name and jobs.agent.runs-on by using the vars context.

You have a GitHub repository that uses GitHub Copilot code review on pull requests.
You plan to add repository-wide code review guidance that will apply to all files.
You need Copilot code review to consistently apply the guidance during pull request reviews.
What should you do?

• A. Create a .github/instructions/copilot.instructions.md file.
• B. Modify the .github/pull_request_template.md file.
• C. Create a .github/copilot-instructions.md file.
• D. Update the instructions in the .github/agents/*.agent.md files.