Your team develops prompts for a customer support GenAI application in Azure AI Foundry, storing the prompt flow in a Git repository. Compliance requires that every prompt version deployed to production be traceable to an exact, immutable point in source history, and that auditors can reproduce the deployed prompt months later. The team currently commits prompt changes to a long-lived 'main' branch and deploys the latest commit. What should you implement to meet the traceability and reproducibility requirements?