Back to Skills

commitlint

majiayu000
Updated Today
1 views
58
9
58
View on GitHub
Metaai

About

This skill validates commit messages against Conventional Commits format using existing commitlint configuration files. It helps developers set up, configure, and troubleshoot commitlint rules, including integration with CI/CD pipelines and pre-commit hooks. It can also extract rules to generate LLM prompts for proper commit message creation.

Quick Install

Claude Code

Recommended
Plugin CommandRecommended
/plugin add https://github.com/majiayu000/claude-skill-registry
Git CloneAlternative
git clone https://github.com/majiayu000/claude-skill-registry.git ~/.claude/skills/commitlint

Copy and paste this command in Claude Code to install this skill

Documentation

Commitlint

Validate commit messages against Conventional Commits format using commitlint configuration and rules.

When to Use This Skill

Use this skill when:

  • Setting up commitlint for a repository
  • Configuring commitlint rules and shareable configurations
  • Integrating commitlint with pre-commit hooks or CI/CD pipelines
  • Extracting rules from commitlint config to generate LLM prompts for commit message generation
  • Validating commit messages programmatically
  • Troubleshooting commitlint configuration or validation errors
  • Understanding commitlint rule syntax and severity levels
  • Detecting commitlint configuration files in a repository

Core Capabilities

Configuration Detection

Commitlint uses cosmiconfig to find configuration files in this priority order:

Dedicated config files:

.commitlintrc
.commitlintrc.json
.commitlintrc.yaml
.commitlintrc.yml
.commitlintrc.js
.commitlintrc.cjs
.commitlintrc.mjs
.commitlintrc.ts
.commitlintrc.cts
commitlint.config.js
commitlint.config.cjs
commitlint.config.mjs
commitlint.config.ts
commitlint.config.cts

Package files:

  • package.json with commitlint field
  • package.yaml (PNPM) with commitlint field

Configuration Formats

JavaScript ES Modules (Recommended):

// commitlint.config.js or commitlint.config.mjs
export default {
  extends: ['@commitlint/config-conventional'],
};

TypeScript:

// commitlint.config.ts
import type { UserConfig } from '@commitlint/types';
import { RuleConfigSeverity } from '@commitlint/types';

const config: UserConfig = {
  extends: ['@commitlint/config-conventional'],
  rules: {
    'type-enum': [RuleConfigSeverity.Error, 'always', ['feat', 'fix', 'docs']],
  },
};

export default config;

JSON:

{
  "extends": ["@commitlint/config-conventional"]
}

YAML:

extends:
  - "@commitlint/config-conventional"

Rule Configuration

Rules are configured as arrays: [level, applicability, value]

Severity Levels:

LevelMeaningTypeScript Enum
0DisabledRuleConfigSeverity.Disabled
1WarningRuleConfigSeverity.Warning
2ErrorRuleConfigSeverity.Error

Applicability:

  • 'always' - Rule must match
  • 'never' - Rule must not match

Example rules:

rules: {
  // Error if type is not in enum
  'type-enum': [2, 'always', ['feat', 'fix', 'docs', 'style', 'refactor', 'perf', 'test', 'build', 'ci', 'chore', 'revert']],

  // Error if type is empty
  'type-empty': [2, 'never'],

  // Error if subject is empty
  'subject-empty': [2, 'never'],

  // Error if header exceeds 100 chars
  'header-max-length': [2, 'always', 100],

  // Error if subject ends with period
  'subject-full-stop': [2, 'never', '.'],

  // Warning if body doesn't have leading blank line
  'body-leading-blank': [1, 'always'],
}

Common Configurations

@commitlint/config-conventional

The most widely used shareable configuration. Default error-level rules:

RuleConfigurationPass ExampleFail Example
type-enum['build', 'chore', 'ci', 'docs', 'feat', 'fix', 'perf', 'refactor', 'revert', 'style', 'test']fix: messagefoo: message
type-case'lowerCase'fix: messageFIX: message
type-emptyneverfix: message: message
subject-casenever + ['sentence-case', 'start-case', 'pascal-case', 'upper-case']fix: some messagefix: Some Message
subject-emptyneverfix: messagefix:
subject-full-stopnever, '.'fix: messagefix: message.
header-max-length100Short headerHeader > 100 chars
body-leading-blankalways (warning)Blank line before bodyNo blank line
body-max-line-length100Lines <= 100 charsLine > 100 chars
footer-leading-blankalways (warning)Blank line before footerNo blank line
footer-max-line-length100Lines <= 100 charsLine > 100 chars

Complete Configuration Schema

// commitlint.config.js
export default {
  // Extend shareable configs (resolved via node resolution)
  extends: ['@commitlint/config-conventional'],

  // Parser preset for parsing commit messages
  parserPreset: 'conventional-changelog-atom',

  // Output formatter
  formatter: '@commitlint/format',

  // Custom rules (override inherited rules)
  rules: {
    'type-enum': [2, 'always', ['feat', 'fix', 'docs', 'style', 'refactor', 'test', 'chore']],
  },

  // Functions that return true to ignore specific commits
  // Merged with default ignores (merge commits, reverts, semver tags)
  ignores: [(commit) => commit.includes('WIP')],

  // Whether to use default ignore patterns
  // Default patterns: 'Merge pull request', 'Revert X', 'v1.2.3', etc.
  defaultIgnores: true,

  // Custom help URL shown on failure
  helpUrl: 'https://example.com/commit-guidelines',

  // Prompt configuration (for @commitlint/cz-commitlint)
  prompt: {
    messages: {},
    questions: {
      type: {
        description: 'Select the type of change:',
      },
    },
  },
};

CLI Usage

Installation

# npm
npm install -D @commitlint/cli @commitlint/config-conventional

# yarn
yarn add -D @commitlint/cli @commitlint/config-conventional

# pnpm
pnpm add -D @commitlint/cli @commitlint/config-conventional

Common Commands

# Lint the last commit
npx commitlint --last

# Lint a range of commits
npx commitlint --from HEAD~5

# Lint from a specific commit
npx commitlint --from abc1234

# Lint message from stdin
echo "feat: add feature" | npx commitlint

# Lint message from file
npx commitlint < .git/COMMIT_EDITMSG

# Lint with edit flag (reads .git/COMMIT_EDITMSG)
npx commitlint --edit

# Lint with custom config path
npx commitlint --config ./custom-commitlint.config.js

# Print resolved config
npx commitlint --print-config

# Strict mode (warnings become exit code 2)
npx commitlint --last --strict

Exit Codes

CodeMeaning
0Success (no errors)
1Lint errors found
2Warnings found (strict mode only)
3Errors found (strict mode)
9Config file missing (with -g flag)

Rule Reference

Type Rules

RuleConditionDefault Applicability
type-enumtype is found in valuealways
type-casetype is in case valuealways, 'lower-case'
type-emptytype is emptynever
type-max-lengthtype has value or less charactersalways, Infinity
type-min-lengthtype has value or more charactersalways, 0

Subject Rules

RuleConditionDefault Applicability
subject-casesubject is in case valuealways
subject-emptysubject is emptynever
subject-full-stopsubject ends with valuenever, '.'
subject-max-lengthsubject has value or less charactersalways, Infinity
subject-min-lengthsubject has value or more charactersalways, 0
subject-exclamation-marksubject has exclamation before :never

Scope Rules

RuleConditionDefault Applicability
scope-enumscope is found in valuealways, []
scope-casescope is in case valuealways, 'lower-case'
scope-emptyscope is emptynever
scope-max-lengthscope has value or less charactersalways, Infinity
scope-min-lengthscope has value or more charactersalways, 0

Header Rules

RuleConditionDefault Applicability
header-caseheader is in case valuealways, 'lower-case'
header-full-stopheader ends with valuenever, '.'
header-max-lengthheader has value or less charactersalways, 72
header-min-lengthheader has value or more charactersalways, 0
header-trimheader has no leading/trailing whitespacealways

Body Rules

RuleConditionDefault Applicability
body-leading-blankbody begins with blank linealways
body-emptybody is emptynever
body-max-lengthbody has value or less charactersalways, Infinity
body-max-line-lengthbody lines have value or less characters (URLs excluded)always, Infinity
body-min-lengthbody has value or more charactersalways, 0
body-casebody is in case valuealways, 'lower-case'
body-full-stopbody ends with valuenever, '.'

Footer Rules

RuleConditionDefault Applicability
footer-leading-blankfooter begins with blank linealways
footer-emptyfooter is emptynever
footer-max-lengthfooter has value or less charactersalways, Infinity
footer-max-line-lengthfooter lines have value or less charactersalways, Infinity
footer-min-lengthfooter has value or more charactersalways, 0

Case Values

For rules that check case (*-case):

[
  'lower-case',    // lowercase
  'upper-case',    // UPPERCASE
  'camel-case',    // camelCase
  'kebab-case',    // kebab-case
  'pascal-case',   // PascalCase
  'sentence-case', // Sentence case
  'snake-case',    // snake_case
  'start-case',    // Start Case
]

Programmatic Usage

Load Configuration

import load from '@commitlint/load';

async function getCommitlintConfig() {
  const config = await load();
  console.log(config.rules);
  return config;
}

Validate Message

import load from '@commitlint/load';
import lint from '@commitlint/lint';

async function validateMessage(message) {
  const config = await load();
  const result = await lint(message, config.rules);

  return {
    valid: result.valid,
    errors: result.errors,
    warnings: result.warnings,
  };
}

LLM Integration Patterns

Extract Rules for Prompt Generation

Generate LLM-friendly constraints from commitlint config:

def extract_rules_for_prompt(config: dict) -> str:
    """Extract commitlint rules into LLM-friendly format."""
    rules = config.get('rules', {})
    prompt_parts = []

    # Extract type-enum if present
    if 'type-enum' in rules:
        level, applicability, types = rules['type-enum']
        if level > 0 and applicability == 'always':
            prompt_parts.append(f"Allowed commit types: {', '.join(types)}")

    # Extract scope-enum if present
    if 'scope-enum' in rules:
        level, applicability, scopes = rules['scope-enum']
        if level > 0 and applicability == 'always' and scopes:
            prompt_parts.append(f"Allowed scopes: {', '.join(scopes)}")

    # Extract header-max-length
    if 'header-max-length' in rules:
        level, applicability, length = rules['header-max-length']
        if level > 0:
            prompt_parts.append(f"Header must be {length} characters or less")

    # Extract subject-case
    if 'subject-case' in rules:
        level, applicability, cases = rules['subject-case']
        if level > 0 and applicability == 'never':
            prompt_parts.append(f"Subject must NOT use: {', '.join(cases)}")

    return '\n'.join(prompt_parts)

Validation Loop

import subprocess

async def validate_with_commitlint(message: str, cwd: Path | None = None) -> tuple[bool, list[str]]:
    """
    Validate commit message with commitlint.

    Args:
        message: The commit message to validate
        cwd: Working directory (defaults to current directory)

    Returns:
        Tuple of (is_valid, error_messages)
    """
    result = subprocess.run(
        ['npx', 'commitlint'],
        input=message,
        capture_output=True,
        text=True,
        cwd=cwd,
    )

    if result.returncode == 0:
        return True, []

    # Parse errors from stderr
    errors = [line.strip() for line in result.stderr.split('\n') if line.strip()]
    return False, errors

Validation loop pattern:

  1. LLM generates commit message based on diff and rules
  2. Run commitlint on generated message
  3. If validation fails, feed errors back to LLM with context
  4. Retry (max 3 times by default)
  5. Return final message (valid or best effort after retries)

Common Issues

Node v24 ESM issues:

Use .mjs extension for ES modules config, or add "type": "module" to package.json

Missing extends:

Config without extends or rules fails with "Please add rules" error. Include at least one:

export default {
  extends: ['@commitlint/config-conventional'],
};

Empty config error:

Config file must have at least extends or rules defined.

Scope enum empty array:

scope-enum with [] passes all scopes. Use specific array to restrict:

rules: {
  'scope-enum': [2, 'always', ['api', 'ui', 'docs']],
}

Subject case trap:

@commitlint/config-conventional uses never with specific cases, meaning those cases are forbidden (not required):

// This forbids sentence-case, start-case, pascal-case, upper-case
'subject-case': [2, 'never', ['sentence-case', 'start-case', 'pascal-case', 'upper-case']]

Pre-commit Integration

For pre-commit hook integration with commitlint, activate the pre-commit skill:

Skill(command: "pre-commit")

Conventional Commits Reference

For Conventional Commits format specification and examples, activate the conventional-commits skill:

Skill(command: "conventional-commits")

References

Official Documentation

Related Specifications

Source Attribution

This skill was created from reference documentation at the commit-polish repository (2025-12-01)

GitHub Repository

majiayu000/claude-skill-registry
Path: skills/commitlint

Related Skills

sglang

Meta

SGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.

View skill

evaluating-llms-harness

Testing

This Claude Skill runs the lm-evaluation-harness to benchmark LLMs across 60+ standardized academic tasks like MMLU and GSM8K. It's designed for developers to compare model quality, track training progress, or report academic results. The tool supports various backends including HuggingFace and vLLM models.

View skill

langchain

Meta

LangChain is a framework for building LLM applications using agents, chains, and RAG pipelines. It supports multiple LLM providers, offers 500+ integrations, and includes features like tool calling and memory management. Use it for rapid prototyping and deploying production systems like chatbots, autonomous agents, and question-answering services.

View skill

llamaguard

Other

LlamaGuard is Meta's 7-8B parameter model for moderating LLM inputs and outputs across six safety categories like violence and hate speech. It offers 94-95% accuracy and can be deployed using vLLM, Hugging Face, or Amazon SageMaker. Use this skill to easily integrate content filtering and safety guardrails into your AI applications.

View skill