Token导航 LogoToken导航

Agent Skill · 研究检索

review-gha-migration(审查全球人道主义协会的移徙)

review-gha-migration 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。

review-gha-migration 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 45,786。

总安装 824周安装 33GitHub Stars 2下载 267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:review-gha-migration(审查全球人道主义协会的移徙)
来源仓库:https://github.com/rwx-cloud/skills
仓库路径:skills/review-gha-migration
安装命令:
npx skills add https://github.com/rwx-cloud/skills --skill review-gha-migration
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/rwx-cloud/skills --skill review-gha-migration

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。

review-gha-migration 用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词或任务场景快速定位候选结果。通过 npx skills add 命令从指定 GitHub 仓库安装使用。安装前需确认权限范围、维护状态,注意是否触发联网或文件操作。

建议结合原始 README 核验具体用法和功能边界。

来源仓库内路径:skills/review-gha-migration

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/rwx-cloud/skills --skill review-gha-migration;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

review-gha-migration 怎么安装?

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

review-gha-migration 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。

review-gha-migration 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。

review-gha-migration 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。

review-gha-migration 用的人多吗?

公开数据是累计安装 824 次,近一周安装 33 次,下载 267 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。

review-gha-migration 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
review-gha-migration

审查全球人道主义协会的移徙 · 本页

8242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

Quick Reference

Fetch and read the cheat sheet before starting. Do NOT use WebFetch — it summarizes and drops critical details. Instead, use Bash to run rwx docs pull and read stdout directly:

rwx docs pull /docs/rwx/migrating/gha-cheat-sheet

Review Procedure

You are reviewing an RWX config that was generated from a GitHub Actions workflow migration. Your job is to catch problems the implementer missed. Approach this as a skeptical reviewer, not as someone defending prior work.

Step 1: Identify the files

The user provides the RWX config path as $ARGUMENTS. If no path is provided, list the files in .rwx/ and ask the user which config to review.

Then locate the corresponding source GitHub Actions workflow. Look for clues:

  • Comments in the RWX config referencing the source file
  • Filename correspondence (e.g., .rwx/ci.yml likely came from .github/workflows/ci.yml)
  • Ask the user if the source is ambiguous

Read both files in full.

Step 2: Inventory the source workflow

Build a checklist from the GitHub Actions workflow. For each item, you will verify it was correctly translated. Extract:

  • Every job and its needs: dependencies
  • Every step within each job (name + what it does)
  • All triggers and their configurations (branches, paths, event types)
  • Every secret referenced
  • Every environment variable (workflow, job, and step level)
  • Matrix strategies and their dimensions
  • Services (databases, caches, etc.)
  • Artifact upload/download pairs
  • Caching steps
  • Composite action references
  • Reusable workflow calls
  • Conditional logic (if: expressions)
  • Timeout and concurrency settings

Step 3: Verify behavioral equivalence

Go through your checklist item by item and verify each one is accounted for in the RWX config. For each item, classify it as:

  • Correct — properly translated to the RWX equivalent
  • Missing — not present in the RWX config and no TODO comment explaining why
  • Wrong — translated but with a semantic difference that changes behavior
  • Degraded — works but lost important properties (e.g., a parallel job became sequential)

Pay special attention to:

  • Steps that were silently dropped during migration
  • if: conditionals that were lost or simplified incorrectly
  • Environment variables that were not carried over
  • Secrets that are referenced but not mapped
  • Matrix dimensions that were flattened or lost
  • Service containers that have no equivalent
  • Artifact passing between jobs that was not preserved

Step 4: Verify RWX optimizations

If you don't already have the reference documentation in context, fetch it now. Do NOT use WebFetch — it summarizes and drops critical details. Instead, use Bash to run rwx docs pull for each doc and read stdout directly. Run both in a single turn as parallel Bash calls:

  • rwx docs pull /docs/rwx/migrating/rwx-reference — full RWX config syntax
  • rwx docs pull /docs/rwx/migrating/gha-reference — GHA-to-RWX concept mapping

If you encounter a question not covered by these references, use rwx docs search "<query>" to find the relevant documentation page, then rwx docs pull the result.

Using the reference documentation, check whether the config takes full advantage of RWX capabilities:

  • DAG structure: Are tasks that can run in parallel actually parallel, or are there unnecessary sequential dependencies?
  • Content-based caching: Are cache keys content-based (not hash-based) where possible?
  • Package substitution: Are there run: steps installing tools that have RWX package equivalents?
  • Task granularity: Could large monolithic tasks be split into parallel subtasks? In particular, flag install tasks that also contain build commands (e.g., npm ci followed by npm run build) — these should be separate tasks for independent cacheability.
  • Trigger optimization: Are triggers using path filters and branch filters effectively?

Step 5: Validate structure

Check the RWX config for structural issues:

  • Required fields are present
  • YAML is well-formed
  • Task ordering reflects the DAG (independent tasks first)
  • No orphaned use references
  • Run rwx lint.rwx/<name>.yml and review the diagnostics

Step 6: Produce the review

Output a structured review with these sections:

Summary: One-line verdict — is this migration correct and ready, or does it need changes?

Issues (if any): A numbered list of problems found, each with:

  • Severity: blocking (must fix before using) or suggestion (improvement opportunity)
  • What's wrong
  • Where in the RWX config it occurs
  • What the fix should be

Checklist: A markdown checklist showing each source workflow item and whether it was correctly translated:

- [x] Job: build — correctly translated as task `build`
- [ ] Job: deploy — missing, no TODO comment
- [x] Trigger: push to main — correctly mapped
- [ ] Secret: DEPLOY_KEY — referenced but not in secrets mapping

Optimization opportunities: Any RWX-specific improvements not yet applied.

If you find blocking issues, offer to fix them directly.

平台分布

Codex

35.87%
按下载量换算96

Claude

27.05%
按下载量换算72

Cursor

20.09%
按下载量换算54

Gemini CLI

9.86%
按下载量换算26

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