Prompt Version Control

Compare these two versions of a prompt and generate a structured changelog entry.

Version {{old_version}} (previous):
{{old_prompt}}

Version {{new_version}} (current):
{{new_prompt}}

Generate:
1. A one-line summary of what changed
2. Detailed diff: what was added, removed, and modified
3. Likely intent behind the changes
4. Potential impact on output quality (positive, negative, or neutral for each change)
5. Risk assessment: could any change cause regressions?

I'm A/B testing two versions of a prompt for {{task_description}}.

Version A:
{{version_a}}

Version B:
{{version_b}}

Evaluation criteria:
{{evaluation_criteria}}

Design an A/B test plan that includes:
1. Sample size recommendation (number of test cases)
2. Test cases that cover normal, edge, and adversarial inputs
3. Scoring rubric for each evaluation criterion (1-5 scale with descriptions)
4. Statistical method for determining a winner
5. A results template I can fill in as I run tests
6. Decision framework: when to pick A, pick B, or iterate further

I recently updated my {{prompt_name}} prompt and I'm seeing unexpected behavior.

Previous version (was working well):
{{previous_version}}

Current version:
{{current_version}}

Observed regression:
{{regression_description}}

Analyze:
1. Which specific changes likely caused the regression
2. Why the change had this unintended effect
3. A fix that addresses the regression without losing the intended improvement
4. Test cases I should run to verify the fix
5. Guidelines for future edits to this prompt to avoid similar regressions

I manage {{prompt_count}} production prompts across {{product_areas}}. I need a rollback strategy for when prompt updates cause issues.

Current process: {{current_process}}

Design a rollback protocol that covers:
1. Detection: how to identify that a prompt update is causing problems (metrics, alerts, user reports)
2. Assessment: severity classification (P0-P3) with response time targets
3. Rollback execution: step-by-step procedure for reverting to the last known good version
4. Communication: templates for notifying stakeholders
5. Post-mortem: analysis template for understanding what went wrong
6. Prevention: checklist to run before deploying future prompt updates

Design a metrics tracking system for prompt version performance.

Prompt name: {{prompt_name}}
Purpose: {{prompt_purpose}}
Current version: {{current_version_number}}

Define metrics across these dimensions:
1. Quality: output accuracy, relevance, completeness (with scoring rubrics)
2. Efficiency: token usage, response time, cost per request
3. Reliability: error rate, refusal rate, format compliance rate
4. User satisfaction: feedback scores, edit rate (how often users modify the output)

For each metric, specify:
- How to measure it
- Baseline from current version
- Target for next version
- Alert threshold that triggers investigation

Create a prompt review process for a team of {{team_size}} people working on {{product_description}}.

Current workflow: {{current_workflow}}
Prompt library size: {{library_size}} prompts

Design a review process that includes:
1. When review is required (which types of changes need approval)
2. Review checklist: what reviewers should evaluate
3. Roles: who reviews what (by prompt category or criticality)
4. Approval workflow: number of approvals needed by prompt tier
5. Automated checks that can run before human review
6. How to handle urgent changes that need to bypass full review
7. Documentation requirements for each approved change