How to set up a linter to enforce clean code automatically

Clean code principles enforced by tooling are worth ten times more than principles enforced by code review. Tools are consistent, fast, and non-confrontational. They catch issues before a human has to.

Set up ESLint with rules that enforce what you've learned:

// .eslintrc.js
module.exports = {
  rules: {
    'max-lines-per-function': ['error', { max: 20 }],
    'max-params': ['error', { max: 3 }],
    'complexity': ['error', { max: 5 }],
    'no-magic-numbers': ['error', { ignore: [0, 1, -1] }],
    'id-length': ['error', { min: 3, exceptions: ['i', 'j', 'id'] }]
  }
};

Add Prettier for formatting so code style is never a code review topic:

// .prettierrc
{
  "semi": true,
  "singleQuote": true,
  "tabWidth": 2,
  "printWidth": 80
}

Install Husky to run lint on every commit:

npx husky-init
npx husky add .husky/pre-commit "npx lint-staged"

// package.json
"lint-staged": {
  "*.js": ["eslint --fix", "prettier --write"]
}

Add lint to your CI pipeline so lint failures block merges to main. This means clean code is a build requirement, not an aspiration.

Don't enable every rule at once. Start with the rules that address the biggest problems in your codebase — usually complexity and function length — and add more over time. A lint setup that produces 800 errors on day one won't be adopted.