CSS Code Quality calculator
March 21, 2026 ยท View on GitHub
Calculate the Code Quality score of your CSS based on a range of different quality guards.
This package analyzes your CSS on a high level and comes up with a score, divided into three areas:
- Maintainability: how difficult is it for someone looking at the CSS from a high level to find the exact spot to fix a bug?
- Complexity: how difficult is it for someone to make a change and be confident that they can make that change without side effects?
- Performance: How likely is the CSS to have a negative impact on performance, based on high-level metrics? (Not including using hardware accelerated transforms and the like, because other tools are more suited for that.)
Installation
npm install @projectwallace/css-code-quality
Usage
import { calculate } from '@projectwallace/css-code-quality'
let css = `my_css { /* ... */ }`
let result = calculate(css)
/*
The result shape looks something like this:
{
"violations": [ ],
"passes": [ ],
"performance": {
"score": 90,
"violations": [ ],
"passes": [ ]
},
"maintainability": {
"score": 100,
"violations": [ ],
"passes": [ ]
},
"complexity": {
"score": 97,
"violations": [ ],
"passes": [ ]
}
}
Each `passes` or `violations` array contains an object that looks like this:
{
"id": "EmptyRules",
"score": 0,
"value": 0
},
{
"id": "AverageSelectorsPerRule",
"score": 0,
"value": 1.5,
"actuals": [
2,
1
]
}
etc. etc.
*/
Metrics
Performance
| Metric | What is tested | Points deducted |
|---|---|---|
Imports | Number of @import rules | 10 per @import |
EmptyRules | Number of empty rules | 1 per empty rule |
SelectorDuplications | Ratio of duplicate selectors | Up to 10 (when uniqueness ratio < 66%) |
DeclarationDuplications | Ratio of duplicate declarations | Up to 10 (when uniqueness ratio < 66%) |
CssSize | Total CSS file size | 5 if size exceeds 200,000 bytes |
TooMuchComments | Total size of comments | 1 per 250 bytes of comments, max 10 |
TooMuchEmbeddedContent | Total size of embedded content (e.g. base64) | 1 per 250 bytes of embedded content, max 20 |
Maintainability
| Metric | What is tested | Points deducted |
|---|---|---|
SourceLinesOfCode | Total source lines of code | 1 per 1,000 lines over 10,000, max 15 |
AverageSelectorsPerRule | Average number of selectors per rule | 5 per selector over 2 (average), max 15 |
AverageDeclarationsPerRule | Average number of declarations per rule | 5 per declaration over 5 (average), max 15 |
MaxSelectorsPerRule | Maximum number of selectors in a single rule | 0.5 per selector over 10, max 15 |
MaxDeclarationsPerRule | Maximum number of declarations in a single rule | 0.5 per declaration over 10, max 15 |
MoreThanMostCommonSelectorsPerRule | Rules that have more selectors than the most common count | 0.01 per rule, max 15 (when > 10% of rules exceed the mode) |
MoreThanMostCommonDeclarationsPerRule | Rules that have more declarations than the most common count | 0.01 per rule, max 15 (when > 10% of rules exceed the mode) |
Complexity
| Metric | What is tested | Points deducted |
|---|---|---|
MoreThanMostCommonSelectorComplexity | Selectors more complex than the most common complexity | 0.01 per selector, max 10 (when > 10% of selectors exceed the mode) |
MoreThanMostCommonSelectorSpecificity | Selectors with higher specificity than the most common specificity | 0.01 per selector, max 10 (when > 10% of selectors exceed the mode) |
MaxSelectorComplexity | Maximum complexity of a single selector | 0.5 per complexity unit over 5, max 5 |
AverageSelectorComplexity | Average selector complexity | 2 per complexity unit over 2 (average), max 10 |
IdSelectorRatio | Ratio of ID selectors | Up to 5 (when ID selector ratio exceeds 1%) |
ImportantRatio | Ratio of !important declarations | Up to 5 (when !important ratio exceeds 1%) |
Related projects
- CSS Analyzer - A CSS Analyzer that goes through your CSS to find all kinds of relevant statistics.
- Wallace CLI - CLI tool for @projectwallace/css-analyzer
- Constyble - CSS Complexity linter
- Color Sorter - Sort CSS colors by hue, saturation, lightness and opacity