Token导航 LogoToken导航

Agent Skill · 研究检索

dead-code-removal(死代码去除)

dead-code-removal 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。

dead-code-removal 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;许可证为 MIT,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 28,058。

总安装 3,280周安装 134GitHub Stars 4下载 1,051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
MIT

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:dead-code-removal(死代码去除)
来源仓库:https://github.com/89jobrien/steve
仓库路径:skills/dead-code-removal
安装命令:
npx skills add https://github.com/89jobrien/steve --skill dead-code-removal
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/89jobrien/steve --skill dead-code-removal

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。

多语言死代码安全移除工具,包含反射调用等边缘情况检测机制。适合生产部署前的bundle瘦身和遗留功能清理等维护场景。自动识别项目结构和入口点,映射关键执行路径保障存活代码。移除前验证序列化、DI容器等隐式引用,防止破坏依赖注入配置。

dead-code-removal 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/dead-code-removal

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/89jobrien/steve --skill dead-code-removal;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

dead-code-removal 怎么安装?

在终端执行 npx skills add https://github.com/89jobrien/steve --skill dead-code-removal。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

dead-code-removal 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。

dead-code-removal 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。

dead-code-removal 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

dead-code-removal 用的人多吗?

公开数据是累计安装 3,280 次,近一周安装 134 次,下载 1,051 次,GitHub 4 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。

dead-code-removal 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。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 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

  • 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。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 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
dead-code-removal

死代码去除 · 本页

3,2804需要联网MIT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

Dead Code Removal

This skill safely identifies and removes unused code across multiple programming languages. It includes comprehensive safety checks to prevent removing code that's actually needed.

When to Use This Skill

  • After refactoring code and removing features
  • Before production deployment to reduce bundle size
  • When cleaning up legacy code
  • When removing deprecated functionality
  • When optimizing codebase size
  • When maintaining code quality standards

What This Skill Does

  1. Language Detection: Identifies project languages and structure
  2. Entry Point Mapping: Maps entry points and critical paths
  3. Dependency Analysis: Builds dependency graphs and usage patterns
  4. Safe Detection: Identifies unused elements with safety checks
  5. Incremental Removal: Removes code incrementally with validation
  6. Backup Creation: Creates backups before making changes

Helper Scripts

This skill includes Python helper scripts in scripts/:

  • find_unused_imports.py: Uses AST parsing to accurately detect unused imports in Python files. Outputs JSON with unused imports and line numbers. python scripts/find_unused_imports.py src/utils.py src/services.py

How to Use

Remove Unused Code

Find and remove unused imports and functions in this project
Clean up dead code in src/ directory, but be conservative

Specific Analysis

Check for unused functions in src/utils/ and remove them safely

Analysis Process

1. Language Detection

Identify Project Type:

  • Python: Look for pyproject.toml, setup.py, requirements.txt
  • JavaScript/TypeScript: Check package.json, tsconfig.json
  • Java: Look for pom.xml, build.gradle
  • Go: Check go.mod
  • Rust: Check Cargo.toml

Detect Entry Points:

  • Python: main.py, __main__.py, app.py, run.py
  • JavaScript: index.js, main.js, server.js, app.js
  • Java: Main.java, *Application.java, *Controller.java
  • Config files: *.config.*, settings.*, setup.*
  • Test files: test_*.py, *.test.js, *.spec.js

2. Build Dependency Graph

Cross-File Dependencies:

  • Track imports and requires
  • Map function/method calls
  • Identify class inheritance
  • Track dynamic usage patterns

Framework Patterns:

  • Preserve framework-specific patterns (Django models, React components, etc.)
  • Check for decorators and annotations
  • Verify entry point registrations

3. Detect Unused Elements

Using Helper Script:

The skill includes a Python helper script for finding unused imports:

# Find unused imports in Python files
python scripts/find_unused_imports.py src/utils.py src/services.py

Unused Imports:

# Python: AST-based analysis
import ast
# Track: Import statements vs actual usage
# Skip: Dynamic imports (importlib, __import__)
// JavaScript: Module analysis
// Track: import/require vs references
// Skip: Dynamic imports, lazy loading

Unused Functions/Classes:

  • Define: All declared functions/classes
  • Reference: Direct calls, inheritance, callbacks
  • Preserve: Entry points, framework hooks, event handlers

4. Safety Checks

Never Remove If:

  • Python: getattr(), eval(), globals() usage detected
  • JavaScript: window[], this[], dynamic import() usage
  • Java: Reflection, annotations (@Component, @Service)
  • Framework patterns: Models, controllers, routes, components
  • Entry points: Main functions, app initialization
  • Test files: All test-related code

Framework Preservation:

Python:

  • Django: Models, migrations, admin registrations
  • Flask: Routes, blueprints, app factories
  • FastAPI: Endpoints, dependencies

JavaScript:

  • React: Components, hooks, context providers
  • Vue: Components, directives, mixins
  • Angular: Decorators, services, modules

Java:

  • Spring: Beans, controllers, repositories
  • JPA: Entities, repositories

5. Safe Removal Process

Backup Creation:

backup_dir="./unused_code_backup_$(date +%Y%m%d_%H%M%S)"
cp -r . "$backup_dir" 2>/dev/null || mkdir -p "$backup_dir" && rsync -a . "$backup_dir"

Incremental Removal:

  1. Create temp file with change
  2. Validate syntax
  3. Run tests if available
  4. Apply or rollback

Validation:

# Python
python -m py_compile file.py
python -m pytest

# JavaScript
npx eslint file.js
npm test

# Java
javac -Xlint file.java
mvn test

Examples

Example 1: Remove Unused Imports

Input: Clean up unused imports in src/utils/helpers.py

Output:

## Dead Code Removal: src/utils/helpers.py

### Analysis
- **File**: `src/utils/helpers.py`
- **Total imports**: 8
- **Unused imports**: 3

### Unused Imports Found
1. `import os` - Line 2 (not used anywhere)
2. `import json` - Line 3 (not used anywhere)
3. `from datetime import timedelta` - Line 5 (not used anywhere)

### Safe to Remove
✅ All 3 imports are safe to remove (no dynamic usage detected)

### Changes

Before

import os import json from datetime import datetime, timedelta import requests

After

from datetime import datetime import requests


### Validation

- ✅ Syntax check passed
- ✅ Tests pass
- ✅ No dynamic imports detected

Example 2: Remove Unused Functions

Input: Find unused functions in src/services/

Output:

## Dead Code Removal: src/services/

### Analysis
- **Files analyzed**: 12
- **Functions found**: 45
- **Unused functions**: 2

### Unused Functions

**1. `formatOldDate()` in `src/services/utils.js`**
- **Location**: Line 34-42
- **Status**: ✅ Safe to remove
- **Reason**: No references found, not exported, not used in tests

**2. `legacyAuth()` in `src/services/auth.js`**
- **Location**: Line 78-95
- **Status**: ⚠️ Preserved (framework pattern)
- **Reason**: Referenced in route configuration (line 12)

### Summary
- **Removed**: 1 function (`formatOldDate`)
- **Preserved**: 1 function (framework usage)
- **Lines removed**: 9
- **Size reduction**: ~300 bytes

Best Practices

Safety Guidelines

Do:

  • Run tests after each removal
  • Preserve framework patterns
  • Check string references in templates
  • Validate syntax continuously
  • Create comprehensive backups
  • Remove incrementally

Don't:

  • Remove without understanding purpose
  • Batch remove without testing
  • Ignore dynamic usage patterns
  • Skip configuration files
  • Remove from migrations
  • Remove exported/public APIs

Detection Patterns

Static Analysis:

  • Use AST parsing for accurate detection
  • Track cross-file references
  • Check for dynamic usage patterns
  • Verify framework-specific patterns

Validation:

  • Always run syntax checks
  • Run tests after removal
  • Verify build still works
  • Check for runtime errors

Reporting

Report Should Include:

  • Files analyzed (count and types)
  • Unused detected (imports, functions, classes)
  • Safely removed (with validation status)
  • Preserved (reason for keeping)
  • Impact metrics (lines removed, size reduction)

Related Use Cases

  • Code cleanup before release
  • Reducing bundle size
  • Removing deprecated code
  • Maintaining code quality
  • Refactoring legacy codebases
  • Optimizing build times

平台分布

Claude Code

27.27%
按下载量换算287

Antigravity

22.07%
按下载量换算232

Gemini CLI

16.62%
按下载量换算175

Codex

12.6%
按下载量换算132

OpenCode

8%
按下载量换算84

windsurf

3.47%
按下载量换算36

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。

继续浏览同类 Skills