Token导航 LogoToken导航TokenDH.com
前端设计需要联网github未标认证来源可访问许可证需确认审计通过

typescript-writing-codeTypeScript writing 代码

Agent Skill

用于辅助文档、README、Markdown、说明文和内容稿件的整理与改写。它适合让 Agent 提炼结构、补齐章节、统一术语、检查链接或把零散材料整理成可读文档。使用时应保留项目已有事实、命令和路径,不要把未确认的信息写成确定结论;涉及对外文案时,还需要控制语气,避免过度营销或夸大能力。

总安装

220

周安装

9

GitHub Stars

2

下载量

71
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-writing-code(TypeScript writing 代码)
来源仓库:https://github.com/ivantorresedge/molcajete.ai
仓库路径:skills/typescript-writing-code
安装命令:
npx skills add https://github.com/ivantorresedge/molcajete.ai --skill typescript-writing-code
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/ivantorresedge/molcajete.ai --skill typescript-writing-code

简介

辅助文档与内容稿件整理,支持 README、Markdown 等格式改写。

  • 可用于提炼结构、统一术语或检查链接完整性。
  • 需保留项目已有事实与路径信息,避免虚构未确认内容。
  • 涉及对外文案时应控制语气,避免过度营销表述。
  • typescript-writing-code 属于前端设计类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

TypeScript Writing Code

Quick reference for writing production-quality TypeScript code. Each section summarizes the key rules — reference files provide full examples and edge cases.

Strict TypeScript Configuration

This project enforces maximum type safety through tsconfig.base.json. Zero any tolerance — no exceptions.

Key Flags

  • strict: true — Enables all strict type-checking options as a group.
  • noImplicitAny: true — Every value must have an explicit or inferable type. No implicit any.
  • strictNullChecks: truenull and undefined are distinct types. Must be handled explicitly.
  • noUncheckedIndexedAccess: true — Array/object index access returns T | undefined. Always check before using.
  • noUnusedLocals: true — Unused variables are compile errors, not warnings.
  • noUnusedParameters: true — Unused function parameters are compile errors.
  • noImplicitReturns: true — Every code path in a function must return a value.
  • isolatedModules: true — Required for Vite/esbuild compatibility. Prevents features that need full-program analysis.

Zero any Policy

// ❌ Wrong — using `any`
function parse(data: any): User {
  return data as User;
}

// ✅ Correct — using `unknown` with narrowing
function parse(data: unknown): User {
  if (!isUser(data)) {
    throw new Error("Invalid user data");
  }
  return data;
}

Safe Indexed Access

With noUncheckedIndexedAccess, array and record access returns T | undefined:

const items = ["a", "b", "c"];
const first = items[0]; // string | undefined — must check

if (first !== undefined) {
  console.log(first.toUpperCase()); // safe
}

const map: Record<string, number> = { a: 1 };
const value = map["b"]; // number | undefined — must check

Catch Blocks

Always type catch variables as unknown:

try {
  await fetchData();
} catch (error: unknown) {
  if (error instanceof Error) {
    console.error(error.message);
  }
  throw error;
}

See references/strict-config.md for the full tsconfig.base.json, flag explanations, and anti-patterns.

Type Safety Patterns

Use TypeScript's type system to catch bugs at compile time, not runtime.

Type Guards

// typeof guard
function formatValue(value: string | number): string {
  if (typeof value === "string") {
    return value.toUpperCase();
  }
  return value.toFixed(2);
}

// Custom type guard with `is` predicate
function isUser(value: unknown): value is User {
  return (
    typeof value === "object" &&
    value !== null &&
    "id" in value &&
    "email" in value
  );
}

Discriminated Unions

Tag union members with a literal type field. Use exhaustive switch for safety:

type Result<T> =
  | { kind: "success"; data: T }
  | { kind: "error"; error: string };

function handle<T>(result: Result<T>): void {
  switch (result.kind) {
    case "success":
      console.log(result.data);
      break;
    case "error":
      console.error(result.error);
      break;
    default: {
      const _exhaustive: never = result;
      throw new Error(`Unhandled case: ${_exhaustive}`);
    }
  }
}

Branded Types

Use branded types for nominal typing when primitive types are too loose:

type UserId = string & { readonly __brand: "UserId" };
type OrderId = string & { readonly __brand: "OrderId" };

function createUserId(id: string): UserId {
  return id as UserId;
}

function getUser(id: UserId): User { /* ... */ }

// getUser(orderId) — compile error, even though both are strings

See references/type-safety.md for generics, utility types, assertion functions, and anti-patterns.

ESM Module Patterns

This project uses ESM ("type": "module") throughout. All packages set "type": "module" in package.json.

Import Rules

  • Named exports preferred — Default exports make refactoring harder and tree-shaking less predictable.
  • import type for types — Biome enforces useImportType and useExportType. Type-only imports are erased at runtime.
  • node: prefix for built-ins — Always use node:path, node:fs, node:crypto.
// ✅ Correct
import { useState, useEffect } from "react";
import type { ReactNode } from "react";
import path from "node:path";

// ❌ Wrong — missing type keyword
import { ReactNode } from "react"; // Biome error: useImportType

Barrel Files

Use barrel files (index.ts) for clean public APIs, but keep them thin:

// components/index.ts
export { Button } from "./Button";
export { Input } from "./Input";
export type { ButtonProps, InputProps } from "./types";

Dynamic Imports

Use dynamic import() for code splitting in routes and heavy modules:

const AdminPanel = lazy(() => import("./pages/AdminPanel"));

See references/esm-modules.md for build output, circular dependency detection, and package.json exports.

Error Handling

Handle errors explicitly. Use result types for expected failures, exceptions for unexpected ones.

Result Type Pattern

type Result<T, E = string> =
  | { ok: true; data: T }
  | { ok: false; error: E };

function createSuccess<T>(data: T): Result<T, never> {
  return { ok: true, data };
}

function createError<E>(error: E): Result<never, E> {
  return { ok: false, error };
}

Discriminated Union Errors

type ApiError =
  | { kind: "not_found"; resource: string }
  | { kind: "validation"; fields: Record<string, string> }
  | { kind: "unauthorized" };

function handleApiError(error: ApiError): string {
  switch (error.kind) {
    case "not_found":
      return `${error.resource} not found`;
    case "validation":
      return Object.values(error.fields).join(", ");
    case "unauthorized":
      return "Please sign in";
  }
}

Try-Catch Rules

// ✅ Correct — catch unknown, narrow, add context
try {
  await api.createUser(data);
} catch (error: unknown) {
  if (error instanceof ApiError) {
    throw new Error(`Failed to create user: ${error.message}`);
  }
  throw error; // Re-throw unexpected errors
}

// ❌ Wrong — catch any, swallow error
try {
  await api.createUser(data);
} catch (e: any) {
  console.log(e.message); // unsafe access
}

See references/error-handling.md for async patterns, retry logic, and when to throw vs return errors.

Biome (Linter & Formatter)

This project uses Biome (v2.3.11) for both linting and formatting. It replaces ESLint and Prettier entirely.

Key Rules

RuleLevelEffect
noExplicitAnyerrorCannot use any type anywhere
noUnusedVariableserrorAll variables must be used
noUnusedImportserrorAll imports must be used
useConsterrorUse const when variable is never reassigned
useImportTypeerrorUse import type for type-only imports
useExportTypeerrorUse export type for type-only exports
noNonNullAssertionwarnAvoid ! postfix operator
a11y recommendedonAccessibility rules for JSX

Formatter Settings

  • Double quotes, semicolons always, trailing commas
  • 2-space indent, 100-character line width
  • Organize imports automatically

Commands

# Check (lint + format check)
biome check .

# Fix (lint fix + format)
biome check --write .

# CI mode (fails on any issue)
biome ci .

Critical Rule

Never add biome-ignore comments. Fix the underlying issue instead of suppressing it. This is a hard project rule — no exceptions.

See references/biome.md for the full biome.json config, VS Code integration, and common fix patterns.

Naming Conventions & Code Quality

Naming Rules

ContextConventionExample
Variables, functionscamelCaseuserName, fetchUser()
Types, interfaces, classesPascalCaseUserProfile, AuthService
ConstantsUPPER_SNAKE_CASEMAX_RETRIES, API_BASE_URL
Fileskebab-caseuser-profile.tsx, auth-service.ts
Component filesPascalCaseUserProfile.tsx, AuthGuard.tsx
Test filesMatch sourceUserProfile.test.tsx
Enum membersPascalCaseUserRole.Admin

Code Quality Rules

  • Zero warnings policy — Treat warnings as errors. Fix them, don't ignore them.
  • No @ts-ignore — Use @ts-expect-error with a comment explaining why, only as a last resort.
  • No type assertions as escape hatchesas unknown as T is a code smell. Refactor instead.
  • Prefer const assertions — Use as const for literal types instead of explicit type annotations.
  • JSDoc for exports — Document exported functions and types with JSDoc. Focus on the "why", not the "what".
/**
 * Encrypts PII fields before database storage.
 * Uses AES-256-GCM with a per-record IV for uniqueness.
 */
export function encryptField(plaintext: string, key: Buffer): EncryptedField {
  // ...
}

Post-Change Verification (MANDATORY)

After every TypeScript code change, run this 4-step verification. No exceptions.

The 4 Steps

# 1. Type-check
pnpm run type-check
# or per-app: pnpm --filter patient type-check

# 2. Lint
pnpm run lint
# or per-app: pnpm --filter patient lint

# 3. Format
pnpm run format
# or per-app: pnpm --filter patient format

# 4. Test
pnpm run test
# or per-app: pnpm --filter patient test

One-Command Verification

# Run all checks for a specific app
pnpm --filter patient validate

Rules

  • All 4 steps must pass before considering a change complete.
  • Fix issues immediately — Don't defer lint warnings or type errors.
  • Never suppress to pass — No @ts-ignore, no biome-ignore, no any casts.

See references/post-change-protocol.md for the full verification workflow, common failures, and troubleshooting.

Reference Files

FileDescription
references/strict-config.mdFull tsconfig.base.json, strict flags, zero-any patterns, anti-patterns
references/type-safety.mdType guards, discriminated unions, branded types, generics, utility types
references/esm-modules.mdESM fundamentals, import/export rules, barrel files, build output
references/error-handling.mdResult types, discriminated union errors, async patterns, retry logic
references/biome.mdFull biome.json config, rules reference, VS Code integration
references/post-change-protocol.md4-step verification workflow, troubleshooting, common failures

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

33.51%
按下载量换算24

Claude

32.74%
按下载量换算23

Cursor

17.41%
按下载量换算12

Gemini CLI

9.19%
按下载量换算7

安全审计

Gen Agent Trust Hub

通过

Socket

通过

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。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