review-pr

name	review-pr
description	Review Pull Request changes against 5 enterprise-ready principles. Evaluates code quality, security, scalability, maintainability, and stakeholder value. Use after CI passes to perform AI-powered code review.
argument-hint	Optional: target branch to compare against (defaults to origin/main)

Review Pull Request

Performs comprehensive AI-powered code review of Pull Request changes against the 5 enterprise-ready principles defined in the project's copilot instructions. This skill assumes CI validation has already passed and focuses on evaluating the quality, design, and architectural aspects of the changes.

When to Use

• Reviewing PR changes before approval
• Getting AI feedback on code quality and design
• Validating adherence to enterprise-ready principles
• After automated CI validations pass successfully
• Before merging changes to main branch

Prerequisites

• Git repository with changes committed
• Current branch differs from target branch
• CI validation pipeline has passed
• Changes follow project naming conventions

The 5 Enterprise-Ready Principles

This skill evaluates changes against these principles in order:

1. Make It Work

• Functional, tested code that meets requirements
• Data flows correctly through medallion architecture (Bronze → Silver → Gold)
• Valid transformations and business logic
• Proper error handling

2. Make It Security

• No hardcoded credentials or sensitive data
• Proper use of environment variables and Key Vault
• Input validation and sanitization
• Principle of least privilege
• No security vulnerabilities in dependencies

3. Make It Scale

• Handles large datasets efficiently
• Proper partition strategies
• Incremental loading patterns (not full refreshes)
• Optimized Spark configurations
• Parallel processing where appropriate

4. Make It Maintainable

• Clear, self-documenting code
• Meaningful variable and function names
• Inline comments for complex logic
• Consistent naming conventions
• Reusable functions (no duplication)
• Comprehensive documentation

5. Delight Stakeholders

• Data quality metrics and monitoring
• Intuitive reports with business context
• Data freshness indicators
• Actionable error messages
• Logging for observability
• Self-service analytics support

Procedure

The skill performs these steps:

1. Compare Branches

• Get current branch name
• Compare with target branch (default: origin/main)
• Collect all changed files with diffs
• Identify artifact types (Notebooks, Lakehouses, Pipelines, etc.)

2. Analyze Changes by Artifact Type

For each changed artifact:

Notebooks:

• Code structure and organization
• Data transformation logic
• Security practices (credentials, SQL injection)
• Performance optimizations
• Documentation quality
• Error handling

Lakehouses:

• Schema design
• Medallion layer organization
• Naming conventions
• Table partitioning strategies

Pipelines:

• Activity configuration
• Error handling and retry logic
• Monitoring and alerting
• Metadata-driven patterns

General Code:

• Code quality and readability
• Security vulnerabilities
• Performance considerations
• Documentation completeness

3. Evaluate Against 5 Principles

For each principle, the AI:

• Reviews relevant code sections
• Identifies strengths and weaknesses
• Provides specific examples from the code
• Suggests improvements with code snippets
• Assigns a rating: ✅ Excellent, ⚠️ Needs Work, ❌ Critical Issues

4. Generate Review Report

Creates a comprehensive review with:

• Executive summary
• Principle-by-principle evaluation
• Overall recommendations
• Code snippets highlighting concerns
• Suggested improvements
• Priority action items

Review Report Format

The skill generates a markdown report structured as:

# Pull Request Review

## Summary
Brief overview of changes and overall assessment.

## Branch Comparison
- **Current Branch**: feature/123_add-validation
- **Target Branch**: origin/main
- **Files Changed**: 5
- **Additions**: 250 lines
- **Deletions**: 50 lines

## Changes by Artifact Type
- Notebooks: 2 modified
- Lakehouses: 1 modified
- Pipelines: 1 new, 1 modified

---

## 1. Make It Work ✅

**Assessment**: Excellent

**Strengths**:
- All functions have proper error handling
- Data validation checks implemented at each stage
- Clear separation between Bronze/Silver/Gold layers

**Code Examples**:
[Specific examples from the code]

---

## 2. Make It Secure ⚠️

**Assessment**: Needs Work

**Concerns**:
- Connection string found in notebook cell 4
- SQL query constructed with string concatenation

**Recommendations**:
[Specific suggestions with code]

---

[Continue for all 5 principles]

---

## Overall Recommendations

1. **Priority 1 (Critical)**: Fix security issues
2. **Priority 2 (High)**: Add optimization for large datasets
3. **Priority 3 (Medium)**: Improve documentation

## Conclusion

[Final summary and approval recommendation]

Scripts

compare-branches.ps1

Compares current branch with target branch and returns detailed diff.

Parameters:

• TargetBranch: Branch to compare against (default: origin/main)
• IncludeDiffs: Include full file diffs (default: true)

Returns:

{
  "currentBranch": "feature/123_add-validation",
  "targetBranch": "origin/main",
  "filesChanged": 5,
  "additions": 250,
  "deletions": 50,
  "files": [
    {
      "path": "NB_1000_BRONZE_Download_Data.Notebook/notebook-content.py",
      "status": "modified",
      "diff": "..."
    }
  ]
}

analyze-artifacts.ps1

Categorizes changed files by Fabric artifact type.

Returns:

{
  "notebooks": ["NB_1000_BRONZE_Download_Data"],
  "lakehouses": ["LH_NYC_Taxi"],
  "pipelines": ["PL_1000_BRONZE_Ingest"],
  "other": ["README.md", "scripts/validate.ps1"]
}

get-file-context.ps1

Extracts relevant context from changed files for AI review.

Parameters:

• FilePath: Path to the file
• MaxLines: Maximum lines to include (default: 500)

Usage Example

As a User Command

@workspace review this PR against the 5 principles

or

@workspace /review-pr

From Another Agent

Please use the review-pr skill to evaluate the current PR changes 
against the enterprise-ready principles defined in our copilot instructions.
Compare against origin/main.

Integration with CI/CD

This skill complements automated CI validation:

1. CI Pipeline (automated):

   • Syntax checks
   • Security scanning
   • Naming convention validation
   • Required file checks

2. PR Review Skill (AI-powered):

   • Code quality assessment
   • Architectural review
   • Design pattern evaluation
   • Business logic validation
   • Stakeholder value assessment

Best Practices

• Run this skill after CI passes (don't repeat automated checks)
• Focus on design, architecture, and business logic
• Provide specific, actionable feedback
• Reference actual code from the PR
• Suggest concrete improvements
• Consider the full context of the project

Limitations

• Cannot execute code or run tests
• Cannot access external systems (databases, APIs)
• Relies on git diff and file content
• Limited to text-based analysis
• Requires good documentation to understand intent

Tips for Better Reviews

1. Clear Commit Messages: Help the AI understand intent
2. Good Documentation: Add comments for complex logic
3. Small PRs: Easier to review thoroughly
4. Meaningful Names: Use descriptive variable/function names
5. Test Coverage: Include test results or descriptions

Output

The skill returns a comprehensive markdown review report that can be:

• Posted as a PR comment
• Saved to a file for reference
• Used to guide code improvements
• Shared with the team for discussion

---

This skill provides AI-powered insights to complement automated validation and human judgment. Always apply critical thinking to AI recommendations.