Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问clear审计通过

law-of-demeter得墨忒耳定律

Agent Skill

law-of-demeter 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

630

周安装

26

GitHub Stars

10

下载量

206
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

3

许可证

MIT

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:law-of-demeter(得墨忒耳定律)
来源仓库:https://github.com/yanko-belov/code-craft
仓库路径:skills/law-of-demeter
安装命令:
npx skills add https://github.com/yanko-belov/code-craft --skill law-of-demeter
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/yanko-belov/code-craft --skill law-of-demeter

简介

law-of-demeter 用于查找、检索和筛选相关信息。

  • 适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。
  • 通过 npx 安装后,结合原始 README 核验具体搜索方式。
  • 安装前需确认权限和维护状态,避免触发不必要的网络请求。
  • 建议根据实际任务场景验证其准确性和覆盖范围。

SKILL.md

Law of Demeter (Don't Talk to Strangers)

Overview

Only talk to your immediate friends, not strangers.

A method should only call methods on: itself, its parameters, objects it creates, or its direct components. Never reach through an object to access another object's internals.

When to Use

  • Accessing nested properties: obj.a.b.c
  • Chaining method calls: obj.getA().getB().getC()
  • Reaching through objects for data
  • Long dot chains in your code

The Iron Rule

NEVER chain through objects. Ask, don't reach.

No exceptions:

  • Not for "it's simpler"
  • Not for "it's just one chain"
  • Not for "the data is there"
  • Not for "fewer lines of code"

Detection: The Chain Smell

If you see multiple dots, you're violating LoD:

// ❌ VIOLATION: Reaching through objects
function getEmployeeCity(company: Company, employeeId: string): string {
  return company.employees
    .find(e => e.id === employeeId)
    ?.address.city;  // Reaching into employee, then into address
}

// More violations:
user.getProfile().getAddress().getZipCode();
order.getCustomer().getPaymentMethod().getLast4();

The Correct Pattern: Ask, Don't Reach

Let objects expose what's needed:

// ✅ CORRECT: Ask the object directly
class Employee {
  constructor(
    private name: string,
    private address: Address
  ) {}

  getCity(): string {
    return this.address.city;  // Employee asks its own address
  }
}

class Company {
  getEmployeeCities(): Map<string, string> {
    return new Map(
      this.employees.map(e => [e.id, e.getCity()])
    );
  }

  getEmployeeCity(employeeId: string): string | undefined {
    return this.employees.find(e => e.id === employeeId)?.getCity();
  }
}

// Usage: Ask company, don't reach through it
const city = company.getEmployeeCity(employeeId);

Why Chains Are Bad

ProblemImpact
Tight couplingCaller knows internal structure
Fragile codeStructure changes break all callers
Hidden dependenciesNot obvious what's needed
Hard to testMust mock entire chain
Null dangerEach . is a potential null

Allowed Method Calls

A method m of class C should only call methods on:

  1. this - C's own methods
  2. Parameters - Objects passed to m
  3. Created objects - Objects m creates
  4. Components - C's direct instance variables
  5. Globals - Accessible global objects (sparingly)
class OrderProcessor {
  constructor(private logger: Logger) {}  // Component

  process(order: Order): Receipt {         // Parameter
    this.validate(order);                  // this
    const receipt = new Receipt(order);    // Created
    this.logger.log('Processed');          // Component
    return receipt;
  }

  // ❌ NOT ALLOWED: order.customer.address.city
  // ✅ ALLOWED: order.getShippingCity()
}

Pressure Resistance Protocol

1. "It's Simpler"

Pressure: "One line with dots is simpler than adding methods"

Response: Simple to write ≠ simple to maintain. Chains create fragile code.

Action: Add methods that expose needed data.

2. "It's Just One Chain"

Pressure: "It's only two dots, not a big deal"

Response: Two dots = two objects you're coupled to. Both can change and break you.

Action: Even short chains should be eliminated.

3. "The Data Is Right There"

Pressure: "The structure has the data, why wrap it?"

Response: Structure changes. Wrapping isolates you from changes.

Action: Ask the owner for the data.

4. "It's Read-Only"

Pressure: "I'm just reading, not modifying"

Response: Reading through chains still couples you to structure.

Action: Ask for what you need.

Red Flags - STOP and Reconsider

If you notice ANY of these, refactor:

  • Multiple dots: a.b.c.d
  • Chained getters: getA().getB().getC()
  • Optional chains: a?.b?.c?.d
  • Null checks for nested access
  • Structure knowledge in calling code
  • Mocking chains in tests

All of these mean: Add a method to ask directly.

Refactoring Chains

// ❌ BEFORE: Chain
const zip = user.getProfile().getAddress().getZipCode();

// ✅ AFTER: Ask
// In User class:
getZipCode(): string {
  return this.profile.getZipCode();
}

// In Profile class:
getZipCode(): string {
  return this.address.zipCode;
}

// Usage:
const zip = user.getZipCode();

Quick Reference

Chain (Bad)Ask (Good)
company.employees[0].address.citycompany.getEmployeeCity(id)
order.customer.paymentMethod.last4order.getPaymentLast4()
user.profile.settings.themeuser.getTheme()
car.engine.fuel.levelcar.getFuelLevel()

Common Rationalizations (All Invalid)

ExcuseReality
"It's simpler"Chains are simpler to write, harder to maintain.
"Just one chain"One chain = multiple couplings.
"Data is right there"Expose it properly through methods.
"It's read-only"Reading chains still couples you.
"Fewer lines"Lines don't matter. Maintainability does.
"It's obvious what it does"Obvious coupling is still coupling.

The Bottom Line

Ask objects for what you need. Don't reach through them.

When you need data from nested objects: add a method on the owner that returns it. Never chain through multiple objects. Each dot is a dependency you're taking on.

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

04

需要参考平台分布和安装热度时

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

补充不同宿主或平台的使用分布数据

能力 5

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

29.82%
按下载量换算61

Claude Code

23.13%
按下载量换算48

windsurf

19.22%
按下载量换算40

Antigravity

12.53%
按下载量换算26

trae

8.51%
按下载量换算18

OpenCode

4.23%
按下载量换算9

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。

来源信息

继续浏览同类 Skills