Add structured checklist and recommendation to AI reviews

[LearningCircuit]

Summary

• Enhances AI reviewer prompt to include structured output format
• Adds "Action Items Checklist" section for important changes only
• Adds "Recommendation" section with clear approve/request changes/uncertain status
• Focuses on avoiding trivial nitpicks and prioritizing actionable feedback

Changes

• Updated system prompt in ai-reviewer.sh to require structured response format
• AI will now end reviews with checkbox checklist of important actionable items
• AI will always provide clear recommendation (Approve/Request Changes/Uncertain)

Test plan

• Test AI review output includes Action Items Checklist
• Test AI review output includes Recommendation section
• Verify checklist focuses on important changes only
• Verify recommendation is clear and actionable

[github-actions]

🤖 AI Code Review

Detailed Review Comments

This diff modifies the system prompt for the AI code reviewer to enforce a more structured and consistent output format. The changes are well-designed and will improve the quality of automated code reviews.

Positive aspects:

• The new structure ensures consistent formatting across all AI-generated reviews
• The "Action Items Checklist" requirement makes it easier to track and address issues
• The "Recommendation" section with standardized options (Approve/Request Changes/Uncertain) provides clear decision-making guidance
• Including an example format helps ensure the AI understands the expected output structure
• The instruction to "ignore trivial nitpicks" helps focus on important issues

No issues found:

• No security vulnerabilities introduced
• No performance impact (only a string literal change)
• No code quality issues
• Follows best practices for prompt engineering

Action Items Checklist

• No action items required - this is a well-implemented improvement

Recommendation

✅ Approve

This change enhances the AI code reviewer's output structure without introducing any issues. The more structured format will make reviews more actionable and consistent, improving the overall code review process.

---

Review by FAIR - needs human verification