mcpField
N
nx

nx

github.com/mindrallyDevelopment287686 installsUpdated 3w ago
Repository

About nx

An Nx development skill for teams organizing and building monorepos. It covers workspace structure, generators, dependency boundaries, task execution, caching, testing, and CI workflows.

Key features

  1. 01Nx workspace and project configuration
  2. 02Generator-based code scaffolding
  3. 03Dependency boundary enforcement
  4. 04Affected task execution and caching
  5. 05Testing and CI workflow guidance

Use cases

  1. 01Structuring apps and libraries in an Nx monorepo
  2. 02Generating consistent projects and components
  3. 03Running tests only for changed projects

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/mindrally/skills --skill nx

After installing, type/nxin Claude Code to confirm it loaded.

Details

From the project README

Nx Monorepo Development

You are an expert in Nx, the smart, fast, and extensible build system for monorepos.

Project Structure

  • Organize projects following Nx conventions:
    • apps/ - Application projects (web apps, APIs, mobile apps)
    • libs/ - Library projects (shared code, features, utilities)
  • Use consistent naming patterns: scope-type-name (e.g., shared-ui-button)
  • Group related libraries under feature folders

Workspace Configuration

Configure nx.json for workspace-wide settings:

{
  "targetDefaults": {
    "build": {
      "dependsOn": ["^build"],
      "cache": true
    },
    "test": {
      "cache": true
    }
  },
  "defaultBase": "main"
}
  • Use project.json for project-specific configuration
  • Define proper tags for enforcing module boundaries

Project Configuration

Each project should have a project.json:

{
  "name": "my-app",
  "sourceRoot": "apps/my-app/src",
  "projectType": "application",
  "tags": ["scope:web", "type:app"],
  "targets": {
    "build": { },
    "serve": { },
    "test": { }
  }
}
  • Define clear project types: application or library
  • Use tags for enforcing dependency constraints

Code Generation

  • Use Nx generators for consistent code scaffolding:
    • nx g @nx/react:app my-app - Generate React application
    • nx g @nx/react:lib my-lib - Generate React library
    • nx g @nx/react:component my-component --project=my-lib - Generate component
  • Create custom generators for project-specific patterns
  • Use --dry-run to preview changes before execution

Module Boundaries

Enforce boundaries using ESLint rules:

{
  "@nx/enforce-module-boundaries": [
    "error",
    {
      "depConstraints": [
        { "sourceTag": "type:app", "onlyDependOnLibsWithTags": ["type:lib", "type:util"] },
        { "sourceTag": "type:lib", "onlyDependOnLibsWithTags": ["type:lib", "type:util"] },
        { "sourceTag": "scope:web", "onlyDependOnLibsWithTags": ["scope:web", "scope:shared"] }
      ]
    }
  ]
}
  • Define clear dependency rules between project types
  • Use scopes to separate domain boundaries

Caching and Performance

  • Enable computation caching for faster builds
  • Configure Nx Cloud for distributed caching and task execution
  • Use affected commands to only run tasks for changed projects:
    • nx affected:build
    • nx affected:test
    • nx affected:lint
  • Define proper inputs and outputs for accurate caching

Task Execution

  • Run tasks with Nx CLI:
    • nx build my-app - Build specific project
    • nx run-many -t build - Build all projects
    • nx affected -t test - Test affected projects
  • Use task pipelines for proper dependency ordering
  • Configure parallel execution for independent tasks

Testing Strategy

  • Use Jest for unit testing with Nx presets
  • Configure Cypress or Playwright for E2E testing
  • Implement component testing for UI libraries
  • Use nx affected:test in CI for efficient test runs

CI/CD Integration

  • Use Nx Cloud for distributed task execution
  • Configure GitHub Actions with Nx:
    - uses: nrwl/nx-set-shas@v4
    - run: nx affected -t lint test build
    
  • Implement proper caching strategies
  • Use nx-cloud record for capturing metrics

Best Practices

  • Keep applications thin; move logic to libraries
  • Create shared utility libraries for common code
  • Use barrel exports (index.ts) for clean imports
  • Implement proper type exports from libraries
  • Document library purposes and public APIs
  • Use Nx Console VS Code extension for visual project management
  • Leverage the project graph for understanding dependencies: nx graph

Comments

Markdown supported