The correct option is Using plain language prompts in Copilot Chat to produce code, reviewing the response, and manually pasting the snippet into the editor.

This approach keeps developers in control of what enters the codebase. Copilot Chat can generate code from natural language and the developer evaluates the response for correctness, security, and style. Manual insertion ensures that only reviewed snippets are adopted and aligns with good engineering practices that favor human oversight and code review.

Invoking an Insert and Run action in the IDE that executes the suggested code immediately without asking for confirmation is inappropriate because it bypasses review and validation. Copilot suggestions should be inspected and tested by the developer before they are executed to avoid introducing errors or vulnerabilities.

Configuring Copilot Chat to deploy generated code straight to Cloud Run without any review is not suitable because Copilot Chat does not perform deployments and automated deployment of unreviewed code violates change control and review practices. Teams should gate deployments behind human approval and established CI and CD workflows.

Allowing Copilot Chat to generate and push commits to the main branch automatically with no developer involvement is unsafe because it sidesteps branch protection, pull requests, and required reviews. Main should be protected and changes should flow through reviewed pull requests to maintain code quality and traceability.

Full AWS Practitioner Certification Question