Token导航 LogoToken导航

Agent Skill · 研究检索

branded-types(品牌类型)

branded-types 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。

branded-types 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 72,083。

总安装 190周安装 8GitHub Stars 2下载 67核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

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:branded-types(品牌类型)
来源仓库:https://github.com/marius-townhouse/effective-typescript-skills
仓库路径:skills/branded-types
安装命令:
npx skills add https://github.com/marius-townhouse/effective-typescript-skills --skill branded-types
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/marius-townhouse/effective-typescript-skills --skill branded-types

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

branded-types 用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词或任务场景快速定位候选结果。通过 npx skills add 命令从指定仓库安装,需结合原始 README 核验具体用法。

安装前建议确认权限范围、维护状态及是否会触发联网、命令执行或文件读写。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/branded-types

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/marius-townhouse/effective-typescript-skills -…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

branded-types 怎么安装?

在终端执行 npx skills add https://github.com/marius-townhouse/effective-typescript-skills --skill bra…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

branded-types 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

branded-types 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

branded-types 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

branded-types 用的人多吗?

公开数据是累计安装 190 次,近一周安装 8 次,下载 67 次,GitHub 2 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

branded-types 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
branded-types

品牌类型 · 本页

1902只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Consider Brands for Nominal Typing

Overview

Add phantom types to distinguish semantically different values.

TypeScript uses structural typing, but sometimes you need nominal typing - values that are distinct because you SAY they are, not because they have different shapes. Brands let you do this without runtime overhead.

When to Use This Skill

  • Distinguishing paths (absolute vs relative)
  • Attaching units to numbers (meters, seconds)
  • Preventing 2D/3D vector mix-ups
  • Marking validated/sanitized strings
  • Creating type-safe identifiers

The Iron Rule

ALWAYS use brands when primitive types have different semantic meanings.

Remember:

  • Brands are phantom types (exist only in type system)
  • No runtime overhead
  • Force explicit conversion/validation
  • Make invalid states unrepresentable

Detection: The "Wrong Primitive" Problem

When different primitives can be confused:

// ❌ Any string can be passed
function readFile(path: string) { ... }

readFile('foo.txt');        // Relative path - might fail
readFile('/home/foo.txt');  // Absolute path - works
// TypeScript can't tell the difference!

The Branding Pattern

Basic Brand Structure

type AbsolutePath = string & { _brand: 'abs' };

function isAbsolutePath(path: string): path is AbsolutePath {
  return path.startsWith('/');
}

function listAbsolutePath(path: AbsolutePath) {
  // Can only be called with validated paths
}

Using Branded Types

function f(path: string) {
  // Must check before using
  if (isAbsolutePath(path)) {
    listAbsolutePath(path);  // OK: path is now AbsolutePath
  }

  listAbsolutePath(path);
  //               ~~~~ Error: string not assignable to AbsolutePath
}

Why Brands Work

You can't actually create a value that is both a string and has a _brand property:

type AbsolutePath = string & { _brand: 'abs' };

// This intersection is "impossible" at runtime
// But TypeScript still uses it for type checking

The only way to get an AbsolutePath is to:

  1. Be given one (from a function that returns it)
  2. Use a type guard to validate and narrow
  3. Use a type assertion (escape hatch)

Common Brand Patterns

Units of Measurement

type Meters = number & { _brand: 'meters' };
type Seconds = number & { _brand: 'seconds' };
type MetersPerSecond = number & { _brand: 'm/s' };

const meters = (m: number) => m as Meters;
const seconds = (s: number) => s as Seconds;

function calculateSpeed(distance: Meters, time: Seconds): MetersPerSecond {
  return (distance / time) as MetersPerSecond;
}

const d = meters(100);
const t = seconds(10);
const speed = calculateSpeed(d, t);  // OK

calculateSpeed(100, 10);  // Error: number not assignable to Meters
calculateSpeed(t, d);     // Error: can't swap distance and time!

Caveat: Arithmetic operations lose the brand:

const doubled = d * 2;
//    ^? const doubled: number  (brand lost)

Validated Strings

type SanitizedHTML = string & { _brand: 'sanitized' };
type UserId = string & { _brand: 'userId' };
type Email = string & { _brand: 'email' };

function sanitize(html: string): SanitizedHTML {
  // Actually sanitize the HTML
  return html.replace(/<script>/g, '') as SanitizedHTML;
}

function setInnerHTML(el: Element, html: SanitizedHTML) {
  el.innerHTML = html;  // Safe: we know it's sanitized
}

// Can't pass unsanitized strings
setInnerHTML(el, '<script>alert("xss")</script>');
//               ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Error!

// Must sanitize first
setInnerHTML(el, sanitize(userInput));  // OK

Type-Safe IDs

type UserId = string & { _brand: 'userId' };
type PostId = string & { _brand: 'postId' };

function getUser(id: UserId): User { ... }
function getPost(id: PostId): Post { ... }

declare const userId: UserId;
declare const postId: PostId;

getUser(userId);  // OK
getUser(postId);  // Error: PostId not assignable to UserId

Sorted Arrays

type SortedList<T> = T[] & { _brand: 'sorted' };

function isSorted<T>(xs: T[]): xs is SortedList<T> {
  for (let i = 0; i < xs.length - 1; i++) {
    if (xs[i] > xs[i + 1]) return false;
  }
  return true;
}

function binarySearch<T>(xs: SortedList<T>, x: T): boolean {
  // Can assume xs is sorted
  let low = 0, high = xs.length - 1;
  while (high >= low) {
    const mid = low + Math.floor((high - low) / 2);
    const v = xs[mid];
    if (v === x) return true;
    [low, high] = x > v ? [mid + 1, high] : [low, mid - 1];
  }
  return false;
}

const nums = [1, 3, 5, 7, 9];
if (isSorted(nums)) {
  binarySearch(nums, 5);  // OK: nums is SortedList<number>
}

Alternative Branding Techniques

Using Unique Symbol (Stronger)

declare const brand: unique symbol;

type Meters = number & { [brand]: 'meters' };

// Can't be faked because brand isn't exported

Using Private Fields in Classes

class ValidatedEmail {
  private readonly _brand!: 'email';
  constructor(public readonly value: string) {
    if (!value.includes('@')) throw new Error('Invalid email');
  }
}

Preventing Vector Mix-ups

interface Vector2D {
  x: number;
  y: number;
  z?: never;  // Explicitly prevent z
}

function norm(v: Vector2D) {
  return Math.sqrt(v.x ** 2 + v.y ** 2);
}

const v3d = { x: 3, y: 4, z: 5 };
norm(v3d);  // Error: z is incompatible with never

Or use brands:

type Vector2D = { x: number; y: number } & { _brand: '2d' };
type Vector3D = { x: number; y: number; z: number } & { _brand: '3d' };

Pressure Resistance Protocol

1. "Just Use Type Aliases"

Pressure: "Type alias is simpler: type UserId = string"

Response: Type aliases don't prevent mixing up different string types.

Action: Use brands when semantic distinction matters.

2. "It's Just Runtime Overhead"

Pressure: "Adding properties to primitives costs memory"

Response: Brands are phantom types - they don't exist at runtime.

Action: Use brands freely; there's no runtime cost.

Red Flags - STOP and Reconsider

  • Multiple string/number types that could be confused
  • Functions that accept "any string" but expect specific formats
  • Validation that happens but isn't tracked in the type system
  • Bugs from swapping similarly-typed arguments

Common Rationalizations (All Invalid)

ExcuseReality
"We'll be careful"Mistakes happen, especially in large codebases
"Type alias is enough"Aliases don't prevent cross-assignment
"Too much ceremony"Prevents bugs that are hard to track down

Quick Reference

// Basic brand pattern
type Brand<T, B extends string> = T & { _brand: B };

type UserId = Brand<string, 'userId'>;
type Meters = Brand<number, 'meters'>;

// Type guard pattern
function isX(val: T): val is BrandedT { ... }

// Factory pattern
const meters = (n: number) => n as Meters;

The Bottom Line

Use brands to give semantic meaning to primitives.

Brands add no runtime overhead but prevent mixing up values that happen to have the same underlying type. Use them for IDs, paths, units, validated strings, and any primitive where semantic distinction matters.

Reference

Based on "Effective TypeScript" by Dan Vanderkam, Item 64: Consider Brands for Nominal Typing.

平台分布

Codex

35.82%
按下载量换算24

Claude

29.41%
按下载量换算20

Cursor

20.88%
按下载量换算14

Gemini CLI

9.33%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills