Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计提醒

typescript-backend-project-setupTypeScript backend project 设置

Agent Skill

typescript-backend-project-setup 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

449

周安装

18

GitHub Stars

305

下载量

145
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:typescript-backend-project-setup(TypeScript backend project 设置)
来源仓库:https://github.com/ntcoding/claude-skillz
仓库路径:skills/typescript-backend-project-setup
安装命令:
npx skills add https://github.com/ntcoding/claude-skillz --skill typescript-backend-project-setup
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/ntcoding/claude-skillz --skill typescript-backend-project-setup

简介

使用 NX monorepo 架构搭建类型安全的 TypeScript 后端工程。

  • 提供严格 linting、100% 测试覆盖率与智能缓存机制。
  • 内置 AI 防护层阻止危险命令误执行。typescript-backend-project-setup 属于研究检索类 Skill,可作为该场景下的辅助能力补充。
  • 适用于大型团队协作与微服务拆分项目初始化。
  • 需预先安装 NX CLI 并理解 workspace 目录结构约定。

SKILL.md

NX Monorepo TypeScript Backend Project Setup

🚨 DO NOT USE PLAN MODE. This skill IS the plan. Follow the steps exactly as written.
⚠️ Check NX docs for latest conventions: https://nx.dev/docs/getting-started/start-new-project NX evolves quickly. Verify these instructions against current NX best practices before use.

Set up NX monorepo for TypeScript backend projects with maximum type safety, strict linting, 100% test coverage, and AI-optimized project structure.

Contents

  1. Phase 1: Define Project Context - Gather requirements
  2. Phase 2: Create NX Workspace - Run NX generator
  3. Phase 3: Install Dependencies - Add plugins and tools
  4. Phase 4: Create Initial Projects - Generate packages and apps
  5. Phase 5: Add Claude Code Integration - Copy AI guardrails and docs
  6. Phase 6: Enforce Strict Standards - Patch configs
  7. Phase 7: Establish Coding Conventions - Add skill content
  8. Phase 8: Activate Git Hooks - Enable pre-commit checks
  9. Phase 9: Verify Setup - Confirm everything works
  10. Phase 10: Document Architecture - Optional interview

When This Activates

  • User requests: "set up typescript backend project", "create backend project", "initialize typescript backend", "create monorepo"
  • Working in an empty or near-empty project folder
  • User asks for backend project scaffolding or boilerplate

Template Location

This skill uses a template located at: typescript-backend-project-setup/template/

The template contains only files NX cannot create: Claude Code integration, documentation structure, and git hooks.

Before starting, ask the user for the full path to the claude-skillz repository so you can locate the template.

Setup Procedure

Phase 1: Define Project Context

Ask the user:

  1. Workspace name - What should this monorepo be called? (lowercase, hyphens ok)
  2. Domain description - Brief description of what this project does
  3. Claude-skillz path - What is the full path to the claude-skillz repository on your system?
  4. Target directory - Where should the project be created? (defaults to current directory)
  5. Initial packages - List any publishable packages to create (e.g., "query, builder, cli")
  6. Initial apps - List any applications to create (e.g., "api, docs")

Phase 2: Create NX Workspace

Priority: Commands > Installs > Patch files (last resort)

Run the NX workspace generator:

npx create-nx-workspace@latest [workspace-name] --preset=ts --pm=pnpm --nxCloud=skip --interactive=false

This creates:

  • nx.json - NX configuration
  • tsconfig.base.json - Base TypeScript config
  • package.json - Root package with NX scripts
  • pnpm-workspace.yaml - Workspace definition
  • .gitignore - Standard ignores

Patch.gitignore - Add test-output:

Add test-output to .gitignore (vitest coverage output):

test-output

Checkpoint: Verify nx report shows NX version.

Phase 3: Install Dependencies

Add testing and code quality tools:

# Add NX plugins
nx add @nx/vitest
nx add @nx/eslint
nx add @nx/node  # Required for creating applications

# Install testing dependencies
pnpm add -D vitest @vitest/coverage-v8

# Install ESLint dependencies (required for strict config)
pnpm add -D typescript-eslint @nx/eslint-plugin eslint-plugin-functional

# Install git hooks
pnpm add -D husky lint-staged

Adding @nx/vitest. Provides integrated test runner with coverage reporting.

Adding @nx/eslint. Provides consistent linting across all projects.

Adding @nx/node. Required for creating Node.js applications.

Adding husky and lint-staged. Provides pre-commit verification gate.

Phase 4: Create Initial Projects

If user specified packages in Phase 1, create them:

# For each package (publishable library with vitest)
nx g @nx/js:library packages/[pkg-name] --publishable --importPath=@[workspace-name]/[pkg-name] --bundler=tsc --unitTestRunner=vitest

If user specified apps in Phase 1, create them:

# For each app (node application - vitest NOT supported, use none)
nx g @nx/node:application apps/[app-name] --unitTestRunner=none

🚨 IMPORTANT:

  • @nx/js:library supports --unitTestRunner=vitest
  • @nx/node:application only supports --unitTestRunner=jest|none (NOT vitest)

After creating projects, run nx sync to update TypeScript project references.

Phase 5: Add Claude Code Integration

Copy template files (only what NX can't create):

Claude Code Integration:

cp -r [claude-skillz-path]/typescript-backend-project-setup/template/CLAUDE.md [target-directory]/
cp -r [claude-skillz-path]/typescript-backend-project-setup/template/AGENTS.md [target-directory]/
cp -r [claude-skillz-path]/typescript-backend-project-setup/template/.claude [target-directory]/

Adding CLAUDE.md. Provides AI context, commands, and project conventions.

Adding .claude/settings.json. Provides permission guardrails and hook configuration.

Adding .claude/hooks/block-dangerous-commands.sh. Prevents destructive git operations (--force, --hard, --no-verify).

Documentation Structure:

cp -r [claude-skillz-path]/typescript-backend-project-setup/template/docs [target-directory]/
cp [claude-skillz-path]/typescript-backend-project-setup/template/repository-setup-checklist.md [target-directory]/

Adding docs/conventions/. Provides coding standards and workflow documentation.

Adding docs/architecture/. Provides system design and domain terminology templates.

Adding docs/project/. Provides project vision and planning templates.

Git Hooks:

cp -r [claude-skillz-path]/typescript-backend-project-setup/template/.husky [target-directory]/

Adding .husky/pre-commit. Provides pre-commit verification (lint, typecheck, test).

Custom ESLint Rules:

cp -r [claude-skillz-path]/typescript-backend-project-setup/template/.eslint-rules [target-directory]/

Adding .eslint-rules/no-generic-names.js. Custom rule that bans generic names (utils, helpers, service, manager) in filenames and class names.

Make scripts executable:

chmod +x [target-directory]/.claude/hooks/block-dangerous-commands.sh

Replace placeholders in copied files:

PlaceholderReplace With
{{WORKSPACE_NAME}}User's workspace name
{{WORKSPACE_DESCRIPTION}}User's domain description
{{DOMAIN_NAME}}User's workspace name (used as context name in glossary)
{{DOMAIN_DESCRIPTION}}User's domain description

Files with placeholders:

  • CLAUDE.md
  • docs/conventions/codebase-structure.md
  • docs/architecture/domain-terminology/contextive/definitions.glossary.yml
  • docs/project/project-overview.md

Phase 6: Enforce Strict Standards

These patches add our strict standards to NX-generated configs.

Patch nx.json - Add lint dependency to build/test:

Add to targetDefaults.build.dependsOn:

"dependsOn": ["lint", "^build"]

Add to targetDefaults.test:

"dependsOn": ["lint"]

This ensures AI gets immediate lint feedback on any change.

Patch tsconfig.base.json - Add strict TypeScript flags:

Add these to compilerOptions:

{
  "noUncheckedIndexedAccess": true,
  "noImplicitOverride": true,
  "noFallthroughCasesInSwitch": true,
  "noPropertyAccessFromIndexSignature": true,
  "noUnusedLocals": true,
  "noUnusedParameters": true,
  "noImplicitReturns": true,
  "exactOptionalPropertyTypes": true,
  "verbatimModuleSyntax": true
}

Patch eslint.config.mjs - Add strict rules:

IMPORTANT: Completely overwrite eslint.config.mjs with this exact content (do not merge, do not patch - replace the entire file):

import nx from '@nx/eslint-plugin';
import tseslint from 'typescript-eslint';
import noGenericNames from './.eslint-rules/no-generic-names.js';

const customRules = {
  plugins: {
    custom: {
      rules: {
        'no-generic-names': noGenericNames,
      },
    },
  },
};

export default tseslint.config(
  ...nx.configs['flat/base'],
  ...nx.configs['flat/typescript'],
  ...nx.configs['flat/javascript'],
  {
    ignores: ['**/dist', '**/out-tsc', '**/node_modules', '**/.nx', '*.config.ts', '*.config.mjs', '*.config.js', 'vitest.workspace.ts'],
  },
  customRules,
  {
    files: ['**/*.ts', '**/*.tsx'],
    rules: {
      // Custom rule: no generic names
      'custom/no-generic-names': 'error',

      // No comments - forces self-documenting code
      'no-warning-comments': 'off',
      'multiline-comment-style': 'off',
      'capitalized-comments': 'off',
      'no-inline-comments': 'error',
      'spaced-comment': 'off',

      // Ban let - use const only
      'no-restricted-syntax': [
        'error',
        {
          selector: 'VariableDeclaration[kind="let"]',
          message: 'Use const. Avoid mutation.',
        },
      ],
      'prefer-const': 'error',
      'no-var': 'error',

      // No any types
      '@typescript-eslint/no-explicit-any': 'error',
      '@typescript-eslint/no-unsafe-assignment': 'error',
      '@typescript-eslint/no-unsafe-member-access': 'error',
      '@typescript-eslint/no-unsafe-call': 'error',
      '@typescript-eslint/no-unsafe-return': 'error',

      // No type assertions - fix the types instead
      '@typescript-eslint/consistent-type-assertions': ['error', { assertionStyle: 'never' }],

      // Ban generic folder imports (not lib - that's NX convention)
      'no-restricted-imports': [
        'error',
        {
          patterns: [
            { group: ['*/utils/*', '*/utils'], message: 'No utils folders. Use domain-specific names.' },
            { group: ['*/helpers/*', '*/helpers'], message: 'No helpers folders. Use domain-specific names.' },
            { group: ['*/common/*', '*/common'], message: 'No common folders. Use domain-specific names.' },
            { group: ['*/shared/*', '*/shared'], message: 'No shared folders. Use domain-specific names.' },
            { group: ['*/core/*', '*/core'], message: 'No core folders. Use domain-specific names.' },
          ],
        },
      ],

      // Complexity limits
      'max-lines': ['error', { max: 400, skipBlankLines: true, skipComments: true }],
      'max-depth': ['error', 3],
      'complexity': ['error', 12],

      // Naming conventions
      '@typescript-eslint/naming-convention': [
        'error',
        {
          selector: 'variable',
          format: ['camelCase'],
        },
        {
          selector: 'variable',
          modifiers: ['const'],
          format: ['camelCase', 'UPPER_CASE'],
        },
        {
          selector: 'function',
          format: ['camelCase'],
        },
        {
          selector: 'parameter',
          format: ['camelCase'],
          leadingUnderscore: 'allow',
        },
        {
          selector: 'typeLike',
          format: ['PascalCase'],
        },
        {
          selector: 'enumMember',
          format: ['PascalCase'],
        },
        {
          selector: 'objectLiteralProperty',
          format: null,
        },
      ],
    },
  },
  {
    files: ['**/*.ts', '**/*.tsx'],
    languageOptions: {
      parserOptions: {
        projectService: true,
        tsconfigRootDir: import.meta.dirname,
      },
    },
  }
);

This enforces:

  • No generic names - Custom rule bans utils, helpers, service, manager, etc. in filenames and class names
  • No let - Only const allowed via no-restricted-syntax
  • No type assertions - Fix the types, don't cast
  • No generic folders - Bans imports from utils/, helpers/, common/, shared/, core/
  • Complexity limits - Max 400 lines, max depth 3, cyclomatic complexity 12
  • No inline comments - Forces self-documenting code
  • No any types - Anywhere
  • Naming conventions - camelCase for variables/functions, PascalCase for types

Patch package.json - Add scripts and lint-staged:

Add these to the root package.json:

{
  "scripts": {
    "build": "nx run-many -t build",
    "test": "nx run-many -t test",
    "lint": "nx run-many -t lint",
    "typecheck": "nx run-many -t typecheck",
    "verify": "nx run-many -t lint,typecheck && nx run-many -t test --coverage",
    "prepare": "husky"
  },
  "lint-staged": {
    "*.ts": ["eslint --fix"]
  }
}

Patch vitest.config.mts files - Add 100% coverage thresholds:

For EACH project created in Phase 4 that has a vitest.config.mts, add thresholds to the coverage block:

coverage: {
  reportsDirectory: './test-output/vitest/coverage',
  provider: 'v8' as const,
  thresholds: {
    lines: 100,
    statements: 100,
    functions: 100,
    branches: 100,
  },
},

This enforces 100% test coverage - tests will FAIL if coverage drops below 100%.

Phase 7: Establish Coding Conventions

Copy content from claude-skillz skills to the docs:

Testing conventions:

  • Read: [claude-skillz-path]/writing-tests/SKILL.md
  • Write to: docs/conventions/testing.md

Adding docs/conventions/testing.md. Provides test naming, assertion patterns, and edge case checklists.

Software design conventions:

  • Read: [claude-skillz-path]/software-design-principles/SKILL.md
  • Write to: docs/conventions/software-design.md

Adding docs/conventions/software-design.md. Provides object calisthenics, fail-fast, and dependency inversion patterns.

Phase 8: Activate Git Hooks

Initialize husky, then overwrite the default pre-commit with our version:

cd [target-directory]
npx husky init
# husky init creates a default pre-commit - overwrite it with ours:
cp [claude-skillz-path]/typescript-backend-project-setup/template/.husky/pre-commit .husky/pre-commit

Phase 9: Verify Setup

# Check NX is working
nx report

# View empty workspace
nx graph

Review the repository-setup-checklist.md and ensure all items are checked.

Checkpoint: All verification commands pass. Ready for first commit.

Phase 10: Document Architecture (Optional)

Offer to interview the user to fill in placeholder content:

Architecture Overview (docs/architecture/overview.md):

  • "What systems does this interact with?"
  • "Who are the primary users?"
  • "What are the main technical components?"

Domain Terminology (docs/architecture/domain-terminology/contextive/definitions.glossary.yml):

  • "What are the key terms in this domain?"
  • "How would you define [term] to a new team member?"

Project Overview (docs/project/project-overview.md):

  • "What problem are you solving?"
  • "Who are the users and what do they need?"
  • "What are your non-negotiable principles?"
  • "What are the project phases?"

Summary

This skill creates an NX monorepo using a command-first approach:

  1. create-nx-workspace for foundation
  2. nx add for plugins
  3. pnpm add for dependencies
  4. Copy template files (only what NX can't create)
  5. Patch configs (last resort)
  6. Verify setup

The result provides:

  • Maximum type safety - strict tsconfig, TypeScript project references
  • Strict linting - no comments, naming conventions, no mutation
  • 100% test coverage - with sensible excludes
  • NX orchestration - caching, affected commands, dependency graph
  • AI guardrails - protected configs, blocked dangerous commands
  • Scalable structure - apps and packages pattern with workspace:* dependencies

For adding projects after setup, see docs/conventions/codebase-structure.md.


Verification Mode

Trigger: "verify typescript setup", "check project setup", "audit monorepo config"

Use this to verify an existing repo has all required configurations.

Step 1: Discover all projects

find . -name "package.json" -not -path "*/node_modules/*" -not -path "*/.nx/*"

Step 2: ESLint Config Verification

Read eslint.config.mjs and verify it contains:

  • Import of .eslint-rules/no-generic-names.js
  • Rule custom/no-generic-names: 'error'
  • Rule no-restricted-syntax with VariableDeclaration[kind="let"] selector
  • Rule @typescript-eslint/consistent-type-assertions with assertionStyle: 'never'
  • Rule no-restricted-imports with patterns for utils, helpers, common, shared, core
  • Rules max-lines: 400, max-depth: 3, complexity: 12

If any missing: List what's missing and offer to fix.

Step 3: Vitest Config Verification

For EACH project discovered in Step 1:

  1. Check if vitest.config.mts exists in that project directory
  2. If exists, read it and verify coverage.thresholds contains:

- lines: 100 - statements: 100 - functions: 100 - branches: 100

If any project missing thresholds: List which projects are non-compliant and offer to fix.

Step 4: Git Hooks Verification

  • .husky/pre-commit contains lint-staged and verify
  • package.json has lint-staged config

Step 5: Gitignore Verification

  • .gitignore contains test-output on its own line

Step 6: Report

Output a summary:

✓ ESLint: All rules configured
✗ Vitest: 2/6 projects missing coverage thresholds
  - apps/eclair
  - apps/docs
✓ Git Hooks: Configured
✓ Gitignore: test-output ignored

If failures: Offer to fix all issues automatically.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.14%
按下载量换算51

Claude

29.02%
按下载量换算42

Cursor

19.38%
按下载量换算28

Gemini CLI

9.8%
按下载量换算14

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills