Token导航 LogoToken导航

Agent Skill · 研究检索

code-simplifier(代码简化)

code-simplifier 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。

code-simplifier 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 50,189。

总安装 624周安装 25GitHub Stars 公开资料未说明下载 202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:code-simplifier(代码简化)
来源仓库:https://github.com/abpai/skills
仓库路径:skills/code-simplifier
安装命令:
npx skills add https://github.com/abpai/skills --skill code-simplifier
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/abpai/skills --skill code-simplifier

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。

code-simplifier 用于审查变更代码并优化其可复用性与可读性,保持功能不变前提下提升质量。聚焦于用户最近修改的文件,遵循项目编码规范,避免破坏接口与外部行为。通过 git diff 识别变更,自动检测问题并提供简化建议,支持按任务范围定制修复范围。

操作涉及代码读取与潜在修改,请确保有足够权限且理解其对原逻辑的影响。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/code-simplifier

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/abpai/skills --skill code-simplifier;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

code-simplifier 怎么安装?

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

code-simplifier 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。

code-simplifier 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。

code-simplifier 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。

code-simplifier 用的人多吗?

公开数据是累计安装 624 次,近一周安装 25 次,下载 202 次。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。

code-simplifier 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
code-simplifier

代码简化 · 本页

624—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

Code Simplifier

Review changed code for reuse, quality, and efficiency, then simplify and fix any issues found.

Core Rules

  1. Preserve functionality exactly.
  2. Keep public behavior, outputs, side effects, and interfaces unchanged.
  3. Follow project-specific coding standards and patterns.
  4. Prefer clarity over compactness.
  5. Avoid clever rewrites that reduce debuggability.

Scope

  • Default to the current task's changed files.
  • Focus on code the user just changed or that you changed in this conversation.
  • Keep fixes behavior-preserving unless the user explicitly asked for broader refactors.

Phase 1: Identify Changes

  1. Run git diff.
  2. If there are staged changes, prefer git diff HEAD so staged and unstaged edits are both visible.
  3. If there are no git changes, inspect the most recently modified files relevant to the task.

Phase 2: Run Three Reviews In Parallel

If subagents are available, launch all three review agents concurrently in a single round and pass each agent the full diff.

If subagents are not available, perform the same three review passes yourself.

Launch all three review agents in the same message. Give each agent the full diff so it can reason about the whole change, not just one file in isolation.

Review 1: Code Reuse Review

Check whether the change reimplements logic that already exists.

  • Search for existing helpers, utilities, and shared abstractions that could replace new code.
  • Look in utility directories, shared modules, and files adjacent to the changed code.
  • Flag new functions that duplicate existing behavior.
  • Flag inline logic that should use an existing utility instead of hand-rolled code.
  • Common duplication targets include string manipulation, path handling, env detection, parsing helpers, and type guards.

Review 2: Code Quality Review

Check for structural and maintainability issues.

  • Redundant state: duplicated state, cached values that could be derived, effects that could be direct calls.
  • Parameter sprawl: new parameters added where restructuring or generalization would be cleaner.
  • Copy-paste variation: near-duplicate blocks that should share an abstraction.
  • Leaky abstractions: exposing internal details or breaking module boundaries.
  • Stringly-typed code: raw strings where constants, unions, or existing types should be used.
  • Unnecessary JSX nesting: wrapper elements that add no layout or semantic value.
  • Unnecessary nesting and branching complexity.
  • Unclear identifiers that obscure intent.
  • Dense logic that should be split into coherent, single-purpose helpers.
  • Fragile one-liners that should be explicit, readable control flow.
  • Unnecessary comments: remove comments that only restate what the code does; keep only non-obvious why.

Review 3: Efficiency Review

Check for unnecessary work and avoidable overhead.

  • Unnecessary work: redundant computation, repeated file reads, duplicate API calls, N+1 patterns.
  • Missed concurrency: independent work done sequentially when it could run in parallel.
  • Hot-path bloat: new blocking work in startup, render, request, or polling paths.
  • Recurring no-op updates: state or store writes that fire even when nothing changed; ensure wrapper updaters preserve no-change signals when a same-reference return should short-circuit downstream updates.
  • Unnecessary existence checks: prefer operating directly and handling errors over TOCTOU pre-checks.
  • Memory issues: unbounded collections, missing cleanup, leaked listeners or subscriptions.
  • Overly broad operations: reading or loading more data than needed.

Phase 3: Simplify and Fix

  1. Aggregate the findings from all three review passes.
  2. Fix each worthwhile issue directly.
  3. If a finding is a false positive or not worth changing, skip it without debate.
  4. Re-check the edited code for correctness and local consistency.
  5. Run available lint, test, or targeted validation when practical.
  6. Briefly summarize what you fixed, what you skipped, and whether the reviewed code was already clean.

Boundaries

Do not:

  • Change business logic or edge-case behavior.
  • Alter API contracts, data formats, or error semantics unless requested.
  • Expand scope beyond recently touched code unless explicitly requested.
  • Over-normalize style at the expense of local codebase conventions.

Decision Heuristics

  • If two versions are equivalent, choose the one a new teammate can understand fastest.
  • Keep useful abstractions; remove only those that add indirection without value.
  • Prefer straightforward flow over reduced line count.
  • Stop when readability gains flatten out.

Output

When finished, briefly summarize:

  • what you fixed
  • what you intentionally skipped
  • whether the reviewed code was already clean

平台分布

Codex

36.77%
按下载量换算74

Claude

29.47%
按下载量换算60

Cursor

19.17%
按下载量换算39

Gemini CLI

10.75%
按下载量换算22

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