code-reviewer

name	code-reviewer
description	Proactive code review skill for comprehensive analysis covering code quality, security, and performance. Specialized for Python codebases with fastMONAI/nbdev/medical imaging focus. This skill should be used when Claude detects code changes, new implementations, or when asked to review code quality. Provides flexible output format adapting to context.

Code Reviewer

Comprehensive code review analysis covering quality, security, and performance with Python/fastMONAI specialization.

When to Use

Proactively activate this skill when:

• Reviewing code changes before commit
• Analyzing new implementations or features
• Detecting potential issues in modified files
• Asked to assess code quality or find problems
• Completing significant code writing tasks (self-review)

Review Process

1. Understand Context First

Before reviewing, establish:

• Purpose of the code (feature, bugfix, refactor)
• Integration points with existing codebase
• Criticality level (production, research, prototype)

2. Apply Review Layers

Review code through three lenses, in order of priority:

Layer 1: Code Quality

• Readability: Clear naming, appropriate comments, logical organization
• Maintainability: Single responsibility, DRY principles, appropriate abstraction level
• Correctness: Logic errors, edge cases, error handling
• Style: PEP 8 compliance, consistent formatting, idiomatic Python

Layer 2: Security

• Input validation: Untrusted data handling, injection risks
• Data exposure: Sensitive data logging, hardcoded secrets
• Resource safety: File operations, network calls, subprocess usage
• Dependencies: Known vulnerabilities, version pinning

Layer 3: Performance

• Algorithmic efficiency: Time/space complexity, unnecessary iterations
• Memory management: Large object handling, memory leaks in loops
• I/O optimization: Batch operations, caching opportunities
• Medical imaging specific: Tensor operations, GPU memory, large volume handling

3. fastMONAI/nbdev Specific Checks

When reviewing fastMONAI code:

• Verify changes are in .ipynb notebooks, not .py files directly
• Check #| export directives are correct
• Ensure #| notest is used appropriately for expensive operations
• Validate padding_mode=0 usage with TorchIO CropOrPad
• Confirm preprocessing consistency between training and inference
• Check MedImage/MedMask type usage for medical data

Reference references/python-patterns.md for detailed Python anti-patterns and best practices.

Output Format Guidelines

Adapt output based on context:

For Small Changes (< 50 lines)

Use inline format with line references:

[file.py:42](file.py#L42) - Issue: Variable `x` shadows outer scope
[file.py:55-60](file.py#L55-L60) - Suggestion: Extract to helper function

For Large Changes or Full Reviews

Use structured report:

## Review Summary
Brief overview of findings

## Critical Issues
Must-fix problems (security, correctness)

## Improvements
Recommended changes (quality, performance)

## Positive Observations
Well-implemented patterns worth noting

## Questions
Clarifications needed from author

For Self-Review (after writing code)

Brief checklist verification:

• Tested edge cases
• Error handling complete
• No obvious security issues
• Performance acceptable

Severity Classification

• Critical: Security vulnerabilities, data loss risks, crash bugs
• High: Logic errors, significant performance issues, maintainability blockers
• Medium: Code smell, minor inefficiencies, style inconsistencies
• Low: Nitpicks, optional improvements, documentation gaps

Review Tone

• Be constructive, not critical
• Explain the "why" behind suggestions
• Acknowledge good patterns alongside issues
• Provide concrete fix suggestions when possible
• Ask questions rather than assume intent