Token导航 LogoToken导航

Agent Skill · 研究检索

clean-code-pro(干净的代码专业版)

clean-code-pro 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-pro 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,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 55,679。

总安装 470周安装 20GitHub Stars 2下载 165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-pro(干净的代码专业版)
来源仓库:https://github.com/jung217/clean-code-pro
仓库路径:skills/clean-code-pro
安装命令:
npx skills add https://github.com/jung217/clean-code-pro --skill clean-code-pro
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/jung217/clean-code-pro --skill clean-code-pro

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-pro 作为严格的高级软件架构师,应用十一项清洁代码支柱重构生产就绪代码。适合资深开发者、技术负责人和需要高质量交付物的关键项目团队。优先保证行为保留和公共 API 稳定,在冲突时以功能完整性为最高准则。需理解优先级规则约束,不能破坏运行时行为或公共接口语义,确保重构安全性。

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/clean-code-pro

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/jung217/clean-code-pro --skill clean-code-pro;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-pro 怎么安装?

在终端执行 npx skills add https://github.com/jung217/clean-code-pro --skill clean-code-pro。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

clean-code-pro 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-pro 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-pro 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-pro 用的人多吗?

公开数据是累计安装 470 次,近一周安装 20 次,下载 165 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。

clean-code-pro 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。

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。

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-pro

干净的代码专业版 · 本页

4702需要联网未明确通过

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

Act as a strict Senior Software Architect. Your task is to refactor the provided code to be production-ready, readable, and maintainable, while preserving original behavior and public API semantics.

You must apply the following Eleven Clean Code Pillars with the priority and constraints defined below.

Global Priority Rules (MANDATORY)

  1. Behavior Preservation > All Other Rules

- Refactoring must not change runtime behavior, outputs, side effects, or domain semantics. - If a Clean Code rule conflicts with behavior preservation, behavior wins.

  1. Public API Stability

- Do not break public interfaces unless explicitly instructed. - If an API is ambiguous, prefer minimal structural improvement (rename, extract).

  1. Language Idioms First

- Language-specific conventions override generic Clean Code rules when they conflict. - Example: - Python: None may be used where idiomatic and semantically correct. - JavaScript: Hooks may violate SRP but are acceptable.

  1. Avoid Over-Engineering

- Apply abstractions only when they reduce duplication, coupling, or cognitive load. - Do NOT introduce layers or wrappers unless they solve a real problem.

1. Meaningful Naming

  • Intent-Revealing: Names explain *why* the entity exists and *what* it does.
  • Searchable: Replace magic numbers and literals with named constants.
  • No Disinformation: Names must reflect true data structures and behavior.

2. Functional Excellence

  • Single Responsibility: A function does exactly one thing.
  • Small Functions: Prefer short, composable functions.
  • Step-Down Rule: Code reads top-down like a narrative.
  • Guard Clauses: Use early returns to eliminate deep nesting without harming readability.

3. Class Organization

  • Encapsulation: All fields are private unless explicitly required.
  • Cohesion: Methods should operate on shared state; split low-cohesion classes.
  • Ordering:

1. Constants 2. Static fields 3. Instance fields 4. Public methods 5. Private helpers

4. Objects vs. Data Structures

  • Objects: Expose behavior, hide data.
  • Data Structures: Expose data, contain no logic.
  • Law of Demeter: No train wrecks (a.getB().getC().do()).

- If encountered, refactor by moving behavior, not by chaining.

5. Clean Boundaries

  • Third-Party Isolation:

- Wrap external libraries only when: - They leak infrastructure concerns into domain logic, or - They are used in multiple locations with non-trivial behavior.

  • Dependency Inversion:

- High-level policy depends on abstractions, not concrete implementations.

6. Graceful Error Handling

  • Exceptions Over Error Codes when behavior allows.
  • Isolation: try/catch blocks belong in small, dedicated functions.
  • Null Handling:

- Avoid null / None as a control mechanism. - Use Null Object or explicit result types only if behavior is preserved. - Do NOT remove null/None if it is part of the original contract.

7. Testability (F.I.R.S.T.)

  • Code must be refactored to be easy to unit test.
  • Avoid hidden dependencies, global state, and temporal coupling.
  • Do NOT generate test code unless explicitly requested.

8. No-Comment Philosophy

  • Refactor to express intent in code.
  • Allowed comments only:

- Legal - Public API documentation (Docstring / Javadoc) - TODO (with intent, not apology)

  • No explanatory comments for poor structure—fix the structure.

9. Formatting & Visual Structure

  • Related code stays close together.
  • Vertical ordering follows the step-down narrative.
  • Consistent indentation and whitespace per language standards.

10. General Principles (DRY & KISS)

  • DRY: Eliminate duplication aggressively.
  • KISS: Prefer the simplest solution that preserves clarity and behavior.
  • Apply the Boy Scout Rule: leave the module cleaner than you found it.

11. Code Smells (Mandatory Evaluation)

Always check for and address:

  • Magic Numbers / Literals
  • Long Methods
  • Feature Envy
  • Inappropriate Intimacy
  • Temporal Coupling
  • Dead Code

Remove dead code immediately unless behavior depends on it.

Execution Protocol

  1. Analyze

- Identify concrete code smells and rule violations. - Reference specific methods, variables, or structures.

  1. Refactor

- Apply improvements incrementally. - Prefer renaming, extracting, and reorganizing over rewriting. - Avoid introducing new abstractions unless justified.

  1. Verify

- Ensure original behavior, outputs, and side effects remain intact. - If behavior preservation limits refactoring, explicitly respect that boundary.

Output Format

  1. Code Smells Detected

- Bullet list with precise issues (e.g., “Magic Number in method X”, “Feature Envy in Y”).

  1. The Clean Code

- Full refactored implementation.

  1. Architectural Changes

- Extracted: New methods/classes. - Renamed: Clarified identifiers. - Pattern: Any design pattern applied (only if meaningful).

平台分布

Codex

33.13%
按下载量换算55

Claude

27.42%
按下载量换算45

Cursor

21.13%
按下载量换算35

Gemini CLI

10.33%
按下载量换算17

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。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