Token导航 LogoToken导航

Agent Skill · 研究检索

shell-scripter(shell 脚本编写器)

shell-scripter 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。

shell-scripter 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,权限为接触敏感数据;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 60,316。

总安装 371周安装 15GitHub Stars 2下载 116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:shell-scripter(shell 脚本编写器)
来源仓库:https://github.com/wyattowalsh/agents
仓库路径:skills/shell-scripter
安装命令:
npx skills add https://github.com/wyattowalsh/agents --skill shell-scripter
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/wyattowalsh/agents --skill shell-scripter

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。

shell-scripter 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。它支持结合来源仓库、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。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。

来源仓库内路径:skills/shell-scripter

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/wyattowalsh/agents --skill shell-scripter;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

    它需要接触敏感数据,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

shell-scripter 怎么安装?

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

shell-scripter 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。

shell-scripter 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

shell-scripter 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。

shell-scripter 用的人多吗?

公开数据是累计安装 371 次,近一周安装 15 次,下载 116 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。

shell-scripter 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

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 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

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 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
shell-scripter

shell 脚本编写器 · 本页

3712接触敏感数据未明确通过

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

Shell Scripter

Generate, review, convert, and lint shell scripts. Makefile and justfile generation. References ShellCheck rule IDs with explanations but does NOT run ShellCheck.

Scope: Shell scripts (bash, zsh, fish, sh/POSIX), Makefiles, justfiles. NOT for Python scripts (use python-conventions), CI/CD pipelines (use devops-engineer), or running/testing scripts.

Dispatch

$ARGUMENTSMode
create <description>Generate a shell script from natural language
review <script or path>Audit for pitfalls, reference ShellCheck rules
convert <script or path> <target>Dialect conversion (bash/zsh/fish)
posix <script or path>POSIX compliance check at pattern level
makefile <tasks>Generate a Makefile from task descriptions
justfile <tasks>Generate a justfile from task descriptions
Natural language about shell scriptingAuto-detect mode from intent
EmptyShow mode menu with examples

Auto-Detection Heuristic

  1. Path to .sh/.bash/.zsh/.fish file + modification verb (review, check, fix, audit, lint) -> review
  2. Path to .sh file + "to zsh/fish/bash" -> convert
  3. Path to .sh file + "posix" or "portable" -> posix
  4. "makefile" or "make targets" in input -> makefile
  5. "justfile" or "just recipes" in input -> justfile
  6. Describes desired script behavior -> create
  7. Ambiguous -> ask which mode

Mode: Create

Generate a shell script from a natural language description.

Generation Process

  1. Determine target dialect (default: bash)
  2. Run uv run python skills/shell-scripter/scripts/dialect-converter.py --list-features <dialect> to confirm available features
  3. Write the script with:

- Proper shebang (env-based like env bash, not hardcoded interpreter paths) - set -euo pipefail for bash (equivalent for other dialects) - Meaningful variable names, quoted expansions - Error handling for external commands - Usage function if script accepts arguments

Validation

  1. Run uv run python skills/shell-scripter/scripts/script-analyzer.py --stdin <<< "$SCRIPT" on the generated script
  2. Fix any issues found, present final script

Mode: Review

Audit a shell script for common pitfalls. Reference ShellCheck rule IDs.

Analysis

  1. Read the target script
  2. Run uv run python skills/shell-scripter/scripts/script-analyzer.py <path>
  3. Parse the JSON output: {shebang, dialect, issues, posix_compatible, complexity_estimate}
  4. For each issue, load references/shellcheck-rules.md to explain the rule ID

Findings Report

  1. Group findings by severity: error > warning > info > style
  2. Present findings with:

- ShellCheck rule ID (e.g., SC2086) - Line number and code snippet - Explanation of WHY it is a problem - Concrete fix

  1. If no issues found, state this explicitly

Severity mapping:

SeverityExamples
errorUnquoted variables in conditionals, syntax errors, command injection
warningMissing error handling, unquoted glob expansions, deprecated syntax
infoSuboptimal patterns, unnecessary subshells, redundant commands
styleInconsistent quoting, missing shellcheck directives, naming

Mode: Convert

Convert shell syntax between bash, zsh, and fish.

  1. Read the source script
  2. Identify source dialect (from shebang or --from flag)
  3. Run uv run python skills/shell-scripter/scripts/dialect-converter.py <path> --from <source> --to <target>
  4. Parse the JSON output: {converted_script, changes, warnings}
  5. Present the converted script with a change summary table
  6. Flag any constructs that have no direct equivalent in the target dialect

Mode: POSIX

Check a script for POSIX compliance at the pattern level.

  1. Read the target script
  2. Run uv run python skills/shell-scripter/scripts/script-analyzer.py <path> --posix
  3. Identify bash-isms: [[]], (()), arrays, local, source, process substitution, {a..z}, $'...'
  4. For each bash-ism, suggest the POSIX equivalent from references/posix-compatibility.md
  5. Report whether the script is POSIX-compatible or list required changes

Mode: Makefile

Generate a Makefile from task descriptions.

  1. Parse task descriptions from $ARGUMENTS
  2. Determine dependencies between tasks
  3. Generate Makefile with:

- .PHONY declarations for non-file targets - .DEFAULT_GOAL - help target using self-documenting pattern (## comments) - Consistent variable naming (UPPER_SNAKE_CASE) - .ONESHELL when multi-line recipes need shared state

  1. Follow conventions from references/makefile-justfile.md

Mode: Justfile

Generate a justfile from task descriptions.

  1. Parse task descriptions from $ARGUMENTS
  2. Generate justfile with:

- Recipe documentation comments - Default recipe (first position or default alias) - Parameter declarations with defaults where sensible - set shell directive if non-default shell needed - set dotenv-load if environment variables are referenced

  1. Follow conventions from references/makefile-justfile.md

Canonical Vocabulary

Use these terms exactly throughout:

TermDefinition
dialectShell language variant: bash, zsh, fish, sh (POSIX)
bash-ismSyntax or feature not in POSIX sh (e.g., arrays, [[]])
shebang#! line specifying the interpreter
SC ruleA ShellCheck rule ID (e.g., SC2086 = unquoted variable)
recipeA justfile target (not "task" or "rule")
targetA Makefile target (not "task" or "recipe")
portableWorks across bash/zsh/sh without modification

Reference Files

Load ONE reference at a time. Do not preload all references into context.

FileContentRead When
references/shellcheck-rules.mdTop 50 ShellCheck rules with severity, explanation, examples, fixesReview mode, explaining SC rule IDs
references/posix-compatibility.mdPOSIX builtins, bash-isms with POSIX equivalents, portability patternsPOSIX mode, create mode with --posix
references/dialect-differences.mdSyntax differences between bash/zsh/fish with conversion recipesConvert mode, cross-dialect questions
references/common-pitfalls.mdUnquoted vars, missing error handling, injection, race conditions, trapsReview mode, create mode best practices
references/makefile-justfile.mdMakefile best practices, justfile syntax and patterns, migration guideMakefile mode, justfile mode
ScriptWhen to Run
scripts/script-analyzer.pyReview and POSIX modes -- static analysis of shell scripts
scripts/dialect-converter.pyConvert mode -- syntax conversion between dialects

Critical Rules

  1. Never claim to run ShellCheck -- reference SC rule IDs and explain them, but analysis is pattern-based
  2. Always use env-based shebangs (e.g., env bash), never hardcoded interpreter paths
  3. Default to set -euo pipefail in generated bash scripts -- omit only with explicit justification
  4. Always quote variable expansions unless splitting is intentionally required (SC2086)
  5. Never generate scripts that use eval unless no alternative exists -- explain the risk
  6. Every generated script must include error handling for external commands
  7. Generated Makefiles must include a .PHONY declaration and a help target
  8. Generated justfiles must have a default recipe and documentation comments
  9. Review findings must include the SC rule ID, line number, and a concrete fix
  10. Never modify the script being reviewed -- review is read-only
  11. Convert mode must warn about constructs with no direct equivalent in the target dialect
  12. POSIX mode must identify every bash-ism and provide a POSIX alternative

平台分布

Codex

32.24%
按下载量换算37

Claude

31.29%
按下载量换算36

Cursor

18.62%
按下载量换算22

Gemini CLI

8.85%
按下载量换算10

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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