quality-review
Review code for quality issues including readability, maintainability, complexity, and adherence to best practices. Use for PR reviews, code audits, identifying technical debt, and ensuring code meets
SKILL.mdskills/old-review/code-review/quality-review/SKILL.md
GitHubCode Quality Review
Evaluate code for readability, maintainability, and adherence to best practices.
Review Focus Areas
Readability
- Clear, descriptive names (variables, functions, classes)
- Appropriate function/method length (< 30 lines preferred)
- Logical code organization
- Consistent formatting and style
- Self-documenting code (minimal comments needed)
Maintainability
- Single Responsibility Principle
- Low coupling between modules
- High cohesion within modules
- No duplicated code (DRY)
- Clear dependency relationships
Complexity
- Cyclomatic complexity (< 10 per function)
- Nesting depth (< 4 levels)
- Number of parameters (< 5 per function)
- Class size (< 300 lines)
- File size (< 500 lines)
Review Checklist
Naming
- Variables describe their content
- Functions describe their action
- Boolean names are questions (is_valid, has_permission)
- No abbreviations except common ones (id, url, etc.)
- Consistent naming conventions throughout
Structure
- Functions do one thing
- No deep nesting (extract to functions)
- Related code is grouped together
- Dependencies are explicit (no globals)
- Error handling is consistent
Code Smells
- No magic numbers/strings (use constants)
- No commented-out code
- No dead code (unreachable paths)
- No overly clever code
- No premature optimization
Issue Severity Levels
CRITICAL - Must fix before merge
- Security vulnerabilities
- Data corruption risks
- Breaking changes without migration
HIGH - Should fix before merge
- Logic errors
- Performance problems
- Missing error handling
MEDIUM - Fix soon after merge
- Code smells
- Minor maintainability issues
- Missing tests for edge cases
LOW - Consider fixing
- Style inconsistencies
- Minor naming improvements
- Documentation gaps
Review Output Format
## Code Review: [File/PR Name]
### Summary
[1-2 sentence overview of the change and overall quality]
### Critical Issues
- **[Location]**: [Issue description]
- Impact: [What could go wrong]
- Suggestion: [How to fix]
### Suggestions
- **[Location]**: [Improvement suggestion]
- Rationale: [Why this matters]
### Positive Notes
- [What was done well]
Common Patterns to Flag
Complexity
# Too complex - extract conditions
if user.is_active and user.has_permission('edit') and not user.is_banned and item.is_editable:
...
# Better
def can_edit(user, item):
return (user.is_active
and user.has_permission('edit')
and not user.is_banned
and item.is_editable)
Magic Values
# Bad
if status == 3:
...
# Good
STATUS_COMPLETED = 3
if status == STATUS_COMPLETED:
...
Deep Nesting
# Bad - deeply nested
def process(items):
for item in items:
if item.is_valid:
if item.needs_processing:
if item.has_data:
# actual logic buried here
# Good - early returns
def process(items):
for item in items:
if not item.is_valid:
continue
if not item.needs_processing:
continue
if not item.has_data:
continue
# actual logic at top level