Token导航 LogoToken导航

Agent Skill · 研究检索

clean-code(干净的代码)

clean-code 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。

clean-code 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,权限为需要联网;许可证信息还不明确,第三方扫描有未通过项,按安装量在研究检索类的 73,675 个 Skill 里排第 43,364。

总安装 989周安装 40GitHub Stars 3下载 310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:clean-code(干净的代码)
来源仓库:https://github.com/fellipeutaka/leon
仓库路径:skills/clean-code
安装命令:
npx skills add https://github.com/fellipeutaka/leon --skill clean-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/fellipeutaka/leon --skill clean-code

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。

clean-code 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。

当前顶部介绍为空,原始 SKILL.md 摘录未提供。clean-code 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/clean-code

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/fellipeutaka/leon --skill clean-code;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

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

clean-code 怎么安装?

在终端执行 npx skills add https://github.com/fellipeutaka/leon --skill clean-code。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

clean-code 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。

clean-code 需要什么权限?

权限级别是需要联网。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。

clean-code 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。

clean-code 用的人多吗?

公开数据是累计安装 989 次,近一周安装 40 次,下载 310 次,GitHub 3 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。

clean-code 安全吗?

第三方扫描有未通过项,风险较高。除非你确认过源码,否则不建议直接安装到常用环境。

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。

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。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 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。

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
clean-code

干净的代码 · 本页

9893需要联网未明确未通过

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

Clean Code

Principles for transforming "code that works" into "code that is clean" — code that can be read, understood, and enhanced by any developer.

"Code is clean if it can be understood easily — by everyone on the team." — Dave Thomas

When to Apply

Reference these guidelines when:

  • Writing new code and choosing names, function signatures, structure
  • Reviewing pull requests for readability and maintainability
  • Refactoring legacy code or reducing complexity
  • Identifying and fixing code smells
  • Improving team code standards

Rule Categories

PriorityCategoryImpact
1NamingHIGH — affects every line of code
2FunctionsHIGH — core unit of abstraction
3Code SmellsHIGH — early detection prevents rot
4FormattingMEDIUM — readability at a glance
5Error HandlingMEDIUM — robustness and clarity
6CommentsMEDIUM — most are avoidable
7Object CalisthenicsASPIRATIONAL — exercises for better OO design

1. Naming

Good names are the single most impactful thing you can do for readability.

Priority order:

  1. Consistency — same concept = same name everywhere
  2. Intent-revealing — name says what it does, not how
  3. Specific — avoid vague names like data, info, manager, handler, utils
  4. Searchable — unique enough to grep
  5. Brief — short but not cryptic
// Bad
const d = new Date();
const arr = users.filter((u) => u.a);
function process(data: any) {}

// Good
const createdAt = new Date();
const activeUsers = users.filter((user) => user.isActive);
function validatePayment(payment: Payment) {}

Conventions:

  • Classes/types: nouns (Customer, OrderRepository). Avoid Manager, Data, Info.
  • Methods/functions: verbs (createOrder, validateEmail, isEligible)
  • Booleans: question form (isActive, hasPermission, canWithdraw)
  • Collections: plural nouns (users, orderItems)

See references/NAMING.md for full guidelines.

2. Functions

// Bad — does too many things, unclear name
function handle(order: Order, sendEmail: boolean, log: boolean) {
  // validate, calculate, save, email, log — all in one
}

// Good — small, single-purpose, descriptive
function validateOrder(order: Order): ValidationResult { ... }
function calculateTotal(items: OrderItem[]): Money { ... }
function saveOrder(order: Order): Promise<void> { ... }

Rules:

  • Small — strive for under 20 lines
  • Do one thing — if you use "and" to describe it, split it
  • One level of abstraction — don't mix business logic with low-level details
  • Few arguments — 0-2 ideal, 3+ warrants a parameter object
  • No side effects — or name them explicitly (saveAndNotify, not save)
  • Command/Query separation — a function either does something or returns something, not both

3. Code Smells

Indicators that code may need refactoring. Not bugs, but design friction.

SmellSymptomQuick Fix
Long Method> 20 lines, multiple concernsExtract methods
Large ClassMany responsibilitiesExtract class (SRP)
Long Parameter List> 3 parametersIntroduce parameter object
Primitive ObsessionStrings/numbers for domain conceptsWrap in value objects
Feature EnvyMethod uses another class's data more than its ownMove method
Data ClumpsSame group of fields appear togetherExtract class
Switch StatementsType-checking switch/if-else across codebaseReplace with polymorphism
Divergent ChangeOne class changed for many reasonsSplit by responsibility
Shotgun SurgeryOne change touches many filesMove related code together
Speculative Generality"Just in case" abstractionsDelete (YAGNI)
Dead CodeUnreachable or unused codeDelete
Message Chainsa.getB().getC().doSomething()Hide delegate (Law of Demeter)

See references/CODE_SMELLS.md for detailed examples and refactoring strategies.

4. Formatting

The Newspaper Metaphor — code should read top-to-bottom like a newspaper article. High-level summary at the top, details below.

class OrderProcessor {
  // Public API first — the "headline"
  process(order: Order): ProcessResult {
    this.validate(order);
    const total = this.calculateTotal(order);
    return this.save(order, total);
  }

  // Supporting methods below, in order of appearance
  private validate(order: Order) { ... }
  private calculateTotal(order: Order): Money { ... }
  private save(order: Order, total: Money): ProcessResult { ... }
}

Rules:

  • Related code stays close together (vertical density)
  • Blank lines between concepts (vertical openness)
  • Variables declared near their usage
  • Caller above callee (stepdown rule)
  • Consistent indentation — non-negotiable

5. Error Handling

  • Exceptions over error codes — keeps happy path clean
  • Don't return null — use undefined, Result types, or throw
  • Don't pass null — leads to defensive checks everywhere
  • Fail fast — validate at boundaries, trust internals
  • Specific exceptions — InsufficientFundsError over generic Error
// Bad — null checks cascade through codebase
function getUser(id: string): User | null {
  return db.find(id);
}
const user = getUser(id);
if (user === null) { ... } // Every caller must check

// Good — throw at boundary, trust within domain
function getUser(id: string): User {
  const user = db.find(id);
  if (!user) throw new UserNotFoundError(id);
  return user;
}

6. Comments

"Don't comment bad code — rewrite it."

Most comments compensate for failure to express intent in code. Prefer self-documenting code over comments.

Good comments:

  • Why something is done (business reason, non-obvious decision)
  • Warnings ("this is slow because X", "order matters here")
  • TODOs with context (link to issue)
  • Legal/license headers
  • Public API docs (JSDoc for libraries)

Bad comments:

  • Restating what the code does (// increment counter)
  • Commented-out code (that's what git is for)
  • Journal/changelog comments
  • Noise (// constructor, // getters)
  • Mandated boilerplate
// Bad — restates the obvious
// Check if user is active
if (user.isActive) { ... }

// Good — explains a non-obvious business rule
// Users who haven't verified email within 30 days are auto-deactivated
// per compliance requirement GDPR-2024-42
if (user.isAutoDeactivated) { ... }

7. Object Calisthenics

Nine exercises from Jeff Bay to improve OO design. Treat these as aspirational targets — strict during practice, pragmatic in production.

#RuleGoal
1One level of indentation per methodExtract methods aggressively
2Don't use elseEarly returns, guard clauses, polymorphism
3Wrap all primitives with domain meaningValue objects (Email, Money, UserId)
4First-class collectionsWrap arrays in domain-specific classes
5One dot per lineLaw of Demeter — talk to friends only
6Don't abbreviateIf the name is too long, the class does too much
7Keep entities smallClasses < 50 lines, methods < 10 lines
8Limit instance variablesStrive for 2-3; forces focused classes
9No getters/settersObjects have behavior, not just data

See references/OBJECT_CALISTHENICS.md for examples of each rule.

Implementation Checklist

Before submitting code:

  • Can a new team member understand this without asking questions?
  • Are names intention-revealing and consistent?
  • Does each function do exactly one thing?
  • Are there any code smells I can fix while I'm here?
  • Are comments explaining *why*, not *what*?
  • Is error handling clean and specific?
  • Did I leave the code better than I found it? (Boy Scout Rule)

平台分布

Codex

37.31%
按下载量换算116

Claude

29.77%
按下载量换算92

Cursor

18.94%
按下载量换算59

Gemini CLI

10.26%
按下载量换算32

安全审计

Gen Agent Trust Hub

未通过

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。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