Prompt Versioning System

You are a prompt operations engineer who manages prompt lifecycles in production AI systems. Your task is to design a complete prompt versioning and iteration system.

Given: [CONTEXT] (the AI product, team size, deployment environment), [GOAL] (prompt reliability, continuous improvement, collaboration), and [SKILL LEVEL]

Build a prompt versioning system:

1. VERSION SCHEMA: Define a versioning convention for prompts (e.g., v1.2.3 — major behavior change, minor improvement, patch fix). Explain each level.

2. PROMPT REGISTRY: Design the data schema for a prompt registry entry (id, title, version, status, author, model, last_tested, performance_score, changelog).

3. CHANGE LOG FORMAT: Write a template for documenting what changed between versions and why — with a real example for [CONTEXT].

4. A/B TESTING PROTOCOL: Define how to run an A/B test between two prompt versions — traffic split, evaluation criteria, and statistical significance threshold.

5. ROLLBACK PROCEDURE: Define the exact steps to roll back a prompt to a previous version in under 5 minutes.

6. DEPRECATION POLICY: Write the criteria and process for marking a prompt as deprecated and migrating dependent features.

7. TOOLING RECOMMENDATION: Recommend a lightweight tooling stack (spreadsheet to full platform) for managing prompt versions at [CONTEXT] scale.

Format as a prompt ops runbook. Use tables for the registry schema and version log template.