Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计提醒

openclaw-expertOpenClaw expert 搜索

Agent Skill

openclaw-expert 用于记录任务执行中的错误、用户纠正、经验和能力缺口,适合在 Codex、Claude、Cursor、Gemini CLI 中希望让 Agent 持续沉淀问题、修正和最佳实践时使用。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。

总安装

3,145

周安装

126

GitHub Stars

公开资料未说明

下载量

1,018
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:openclaw-expert(OpenClaw expert 搜索)
来源仓库:https://github.com/fanthus/agent-skills
仓库路径:skills/openclaw-expert
安装命令:
npx skills add https://github.com/fanthus/agent-skills --skill openclaw-expert
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/fanthus/agent-skills --skill openclaw-expert

简介

openclaw-expert 检索 OpenClaw 官方文档与 GitHub 资源解答技术疑问与配置难题。

  • 按问题类型路由至对应知识库章节,返回带引用链接的结构化答案便于溯源查证。
  • 覆盖安装指南、API 用法与故障排除全流程,帮助开发者快速跨越学习曲线投入使用。
  • 所有信息均来自公开渠道,不包含私有部署环境下的特殊配置与内部最佳实践分享。
  • 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。

SKILL.md

OpenClaw Learning Expert

This skill helps answer questions about OpenClaw by retrieving information from official documentation and the GitHub repository, then providing comprehensive answers with source links.

Workflow

When the user asks an OpenClaw-related question:

Step 1: Identify the Question Type

Categorize the question to determine the best sources:

  • Getting Started/Installation → Documentation: /start/getting-started
  • Concepts/Architecture → Documentation: /concepts/ sections
  • API Usage → Documentation: /api/ + GitHub examples
  • Configuration → Documentation: /guides/configuration
  • Troubleshooting → GitHub Issues + Documentation
  • Examples/Implementations → GitHub /examples directory
  • Advanced/Source Code → GitHub repository source code

Step 2: Fetch Relevant Documentation

Use web_fetch tool to retrieve content from:

  1. Primary source: Official documentation at https://docs.openclaw.ai/

- Start with the most relevant documentation page based on the question type - Common pages: /start/getting-started, /concepts/, /api/, /guides/

  1. Secondary source: GitHub repository at https://github.com/openclaw/openclaw

- For code examples, implementation details, or when docs need clarification - Check README.md, examples directory, or source code as needed

Important: Always fetch the actual pages rather than guessing content, as OpenClaw is actively developed and documentation changes frequently.

Step 3: Synthesize Information

After retrieving documentation:

  1. Extract relevant information that answers the user's question
  2. Organize the answer in a clear, logical structure:

- Start with a direct answer to the question - Provide necessary context or explanation - Include code examples if relevant - Note any caveats or best practices

  1. Cite sources by including the specific documentation URLs used

Step 4: Present the Answer

Format the response as follows:

[Direct answer to the question]

[Explanation and details]

[Code examples if applicable]

**Sources:**
- [Specific page title]: [Full URL to the documentation page]
- [Another source if used]: [Full URL]

Example response structure:

OpenClaw uses a declarative configuration approach for defining workflows.

To configure a workflow, you create a YAML file that specifies...

Example:

workflow: name: example steps: - action: process


**Sources:**

- Getting Started Guide: [https://docs.openclaw.ai/start/getting-started](https://docs.openclaw.ai/start/getting-started)
- Configuration Reference: [https://docs.openclaw.ai/guides/configuration](https://docs.openclaw.ai/guides/configuration)

Best Practices

  1. Always fetch current documentation - Don't rely on cached knowledge
  2. Provide specific URLs - Include the exact page where information was found
  3. Include code examples - When available in the documentation, include them
  4. Be comprehensive - Cover edge cases and common pitfalls mentioned in docs
  5. Link to GitHub for implementation - When users need to see source code or examples
  6. Check multiple sources - If documentation is unclear, cross-reference with GitHub
  7. Note version information - If the documentation mentions specific versions, include that context

Handling Common Scenarios

Question Not Directly Answered in Docs

  1. Search GitHub Issues for similar questions
  2. Check GitHub Discussions
  3. Examine source code or examples for implementation patterns
  4. Provide best available information with caveats

Multiple Possible Answers

  1. Present all relevant approaches found in documentation
  2. Note recommended approach if docs specify one
  3. Explain trade-offs when applicable

Outdated or Conflicting Information

  1. Prioritize official documentation over GitHub README
  2. Note any conflicts found between sources
  3. Suggest checking GitHub Issues for latest updates
  4. Provide the most recent information available

Reference Files

  • references/documentation_guide.md - Overview of documentation structure and search strategies (consult when unsure where to find specific information)

Tools to Use

  • web_fetch - Primary tool for retrieving documentation pages
  • web_search - For finding specific pages or GitHub issues when exact URL is unknown

Notes

  • OpenClaw is actively developed - always fetch fresh documentation
  • User's questions may be in Chinese or English - respond in the same language
  • Include both Chinese and English technical terms when appropriate
  • Always verify URLs work before including in response

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

37.89%
按下载量换算386

Claude

29.51%
按下载量换算300

Cursor

20.32%
按下载量换算207

Gemini CLI

8.58%
按下载量换算87

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills