The Coding Agent is an AI execution agent on GitHub that understands natural-language requirements, plans the steps, and performs concrete actions inside a repository. Instead of only generating snippets, it acts like a collaborative developer you can delegate tasks to. GitHub Copilot’s Agent mode embodies this idea: you specify a goal, the agent reads the repository, creates an action plan, and can automatically modify files, run tests, commit changes, and even open pull requests.
Key capabilities:
- Intent understanding: turn natural-language or semi-structured descriptions into actionable subtasks.
- Context awareness: inspect files, dependencies, conventions, and existing implementations to keep changes grounded in the project.
- Deliverables: produce patches, commits, branches, and pull requests as reviewable increments.
- Explainability: summarize what it plans to do, why, and what comes next.
We will ask GitHub Copilot to:
- Resolve the two issues created in Lab 3 using the coding agent.
- Node.js: >= 22.0.0
- npm: >= 10.0.0
- VS Code: latest version
- GitHub Copilot: signed in
Use the coding agent to complete the two issues created in Lab 3.
-
Assign the first issue to Copilot On GitHub, open the first issue and click Assign to Copilot.
-
Assign the second issue Repeat the process for the second issue.
-
Review the pull request On the issue page, Copilot will spin up a pull request to address the issue.
-
Monitor progress From the issue page, open the pull request and click View session to see Copilot’s progress. Note: depending on complexity, the agent may take several minutes. You can work on other tasks while it runs.
-
Review the changes After a few minutes, check the Files changed tab to inspect the modifications. If something needs adjustment, @mention Copilot for further changes.
-
Merge the pull request Once satisfied, click Merge pull request to merge into
main.
- After syncing the latest code locally, the project runs correctly.