improve
Analyze code and provide specific improvement suggestions with code examples.
When & Why to Use This Skill
This Claude skill functions as an automated code auditor that analyzes source code to provide actionable improvement suggestions. It focuses on four critical pillars: performance optimization, code readability, security hardening, and adherence to industry best practices. By delivering side-by-side code examples and clear justifications, it helps developers refactor complex logic, eliminate technical debt, and ensure high-quality software delivery.
Use Cases
- Automated Pull Request Reviews: Streamline the development workflow by automatically scanning code changes for performance bottlenecks and style inconsistencies before human review.
- Legacy Code Refactoring: Analyze older modules to identify deeply nested logic or outdated patterns and replace them with modern, readable, and maintainable language idioms.
- Security Vulnerability Detection: Proactively identify risks such as improper input validation or insecure API usage within the codebase to prevent potential exploits.
- Performance Tuning: Detect inefficient algorithms and redundant computations, providing optimized alternatives like list comprehensions or batch processing to improve execution speed.
| name | improve |
|---|---|
| description | Analyze code and provide specific improvement suggestions with code examples. |
| version | 1.0.0 |
Code Improve Skill
Task
Analyze code changes and provide specific improvement suggestions with code examples.
Improvement Categories
Performance
- Algorithm optimization
- Reduce unnecessary computation
- Cache utilization
- Batch processing
Readability
- Naming improvements
- Code simplification
- Extract functions/methods
- Reduce nesting
Security
- Input validation
- Secure API usage
- Sensitive data handling
Best Practice
- Language idioms
- Design patterns
- Error handling
- Type safety
Output Format
suggestions:
- relevant_file: "src/example.py"
language: "python"
relevant_lines_start: 10
relevant_lines_end: 15
label: "performance"
severity: "medium"
one_sentence_summary: "Use list comprehension instead of loop"
suggestion_content: |
Current code uses for loop to build list, can be simplified with list comprehension.
existing_code: |
result = []
for i in range(100):
result.append(i * 2)
improved_code: |
result = [i * 2 for i in range(100)]
Principles
- Only suggest improvements with real value
- existing_code must be actual code from the PR
- improved_code must be directly usable as replacement
- Prioritize high-impact improvements