Token导航 LogoToken导航

Agent Skill · 研究检索

global-search(全局搜索)

global-search 用于查找、检索和筛选相关信息,适合在 OpenClaw 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。

global-search 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为接触敏感数据;许可证为 MIT-0,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 11,575。

总安装 11,136周安装 464GitHub Stars 8下载 3,712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-0
版本
1.1.9

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:global-search(全局搜索)
来源仓库:https://github.com/yangzhichao814-cell/global-search
安装命令:
openclaw skills install global-search
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 OpenClaw 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。

ClawHubOpenClaw
openclaw skills install global-search

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。

融合多源网络信息,提供全面且相关的搜索结果聚合。适用于深度调研、问题解答与知识检索应用场景。通过 clawhub 安装,需确认搜索引擎 API 调用权限。支持自然语言查询与结果排序优化。global-search 属于研究检索类 Skill,可作为该场景下的辅助能力补充。建议结合可信度评分筛选高价值信息来源。

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 OpenClaw;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 openclaw skills install global-search;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

global-search 怎么安装?

在终端执行 openclaw skills install global-search。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

global-search 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。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。

global-search 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。

global-search 是免费的吗?

来源标注的许可证是 MIT-0。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。

global-search 用的人多吗?

公开数据是累计安装 11,136 次,近一周安装 464 次,下载 3,712 次,GitHub 8 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。

global-search 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

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,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

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 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 OpenClaw,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

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

全局搜索 · 本页

11,1368接触敏感数据MIT-0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

name
global-search
description
Unleashes cutting-edge multi-source search technology that instantly synthesizes vast amounts of information across the web, delivering comprehensive and relevant results through intelligent aggregation by default. Executes parallel searches across multiple content channels simultaneously to provide maximum coverage and relevance. Offers superior search efficiency with a single query. Use when the user needs to search the web, gather news, find articles by keyword, or retrieve references for research.
metadata
openclaw
emoji
🔍
required_env_vars
homepage
https://clb.ciglobal.cn
publisher
clb.ciglobal.cn

Global Search

This skill performs multi-source web search through an external service and can aggregate results across multiple sources when the user explicitly asks for comprehensive coverage.

⚠️ Privacy and Security Notice

Important: This skill sends the user’s search query to an external web search service at https://clb.ciglobal.cn/web_search. Please be aware:

  • Data Transmission: Queries are transmitted to a third-party service and may be associated with your account
  • API Key Required: The only runtime credential is GLOBAL_SEARCH_API_KEY
  • Credential Storage: Store the API key securely using environment variables or your system's credential manager. Do not hardcode it in scripts
  • User Consent: If a query may contain sensitive, personal, confidential, or internal information, ask the user to confirm before sending it to the external service
  • Sensitive Query Filter: Do not send passwords, tokens, session cookies, private keys, personal identifiers, internal-only content, or other confidential data
  • Data Minimization: Prefer shortening, generalizing, or redacting unnecessary sensitive details before searching
  • Scope Limitation: Use this skill only for legitimate search purposes
  • Provider Verification: The provider clb.ciglobal.cn is an external third-party service. Its trustworthiness, retention policy, and privacy practices should be verified by the skill publisher before distribution; users should not assume it has been independently vetted
  • Account Association: The API key associates search activity with your account, so search history may be logged or retained by the provider

Credential Configuration

Required Environment Variable:

  • GLOBAL_SEARCH_API_KEY: API key obtained from https://clb.ciglobal.cn/apiKey/login

Recommended Setup:

# Set environment variable (Linux/Mac)
export GLOBAL_SEARCH_API_KEY="your_api_key_here"

# Set environment variable (Windows PowerShell)
$env:GLOBAL_SEARCH_API_KEY="your_api_key_here"

Alternatively, configure the API key in your agent's credential manager or secrets storage.

When to Use

Apply this skill when the user:

  • Asks to search the web or gather information online
  • Needs news articles or references by keyword
  • Wants to retrieve content from diverse online sources
  • Requires comprehensive, real-time web search with maximum coverage

Use comprehensive search only when the user explicitly requests broad multi-source coverage. For simple lookups, prefer the minimum necessary query scope.

API Overview

Endpoint: POST /web_search

Base URL: https://clb.ciglobal.cn

Authentication: Required header X-API-Key. Get your API key at https://clb.ciglobal.cn/apiKey/login

Note: This skill sends user queries to an external web search service. Please ensure you have obtained proper authorization before using this service.

Request

For comprehensive search (default behavior), the skill will use the script from overall.md to perform 4 parallel API calls automatically:

  • Call 1: search_source=baidu_search, mode=network (百度新闻/资讯)
  • Call 2: search_source=google_search, mode=network (谷歌新闻/资讯)
  • Call 3: search_source=baidu_search_ai, mode=network (百度 AI 搜索)
  • Call 4: mode=warehouse (Elasticsearch 索引库,忽略 search_source)

Headers

HeaderRequiredDescription
X-API-KeyYesAPI key for authentication. Get your key at https://clb.ciglobal.cn/apiKey/login
Content-TypeYesapplication/x-www-form-urlencoded (form data)

Form Parameters

ParameterTypeRequiredDefaultDescription
keywordstringYes-Search keyword(s),多个关键词用空格分隔
search_sourcestringNo-Engine: baidu_search, google_search, baidu_search_ai. Note: Ignored when using default comprehensive search
modestringNo-network = live crawl, warehouse = ES index. Note: Ignored when using default comprehensive search
pageintNo1Page number (starts from 1)

Comprehensive Search (All Sources)

当用户明确要求进行全面搜索(即同时搜索所有可用来源)时,使用 overall.md 中的脚本进行搜索,而不是使用下面的示例代码。

When the user explicitly wants to search across ALL available sources simultaneously (comprehensive search), you should:

Example Implementation (Python asyncio):

import aiohttp
import asyncio
import os

API_URL = "https://clb.ciglobal.cn/web_search"
# Get API key from environment variable (recommended)
API_KEY = os.environ.get("GLOBAL_SEARCH_API_KEY", "your_api_key_here")

if API_KEY == "your_api_key_here":
    raise ValueError("Please set GLOBAL_SEARCH_API_KEY environment variable or provide your API key")

headers = {
    "X-API-Key": API_KEY,
    "Content-Type": "application/x-www-form-urlencoded"
}

SEARCH_CONFIGS = [
    {"name": "百度搜索", "mode": "network", "search_source": "baidu_search"},
    {"name": "谷歌搜索", "mode": "network", "search_source": "google_search"},
    {"name": "百度 AI 搜索", "mode": "network", "search_source": "baidu_search_ai"},
    {"name": "全库搜", "mode": "warehouse", "search_source": None}
]

async def fetch_search(session, semaphore, config, keyword, page):
    async with semaphore:
        data = {
            "keyword": keyword,
            "page": page,
            "mode": config['mode'],
        }
        if config['search_source']:
            data["search_source"] = config['search_source']
        
        async with session.post(API_URL, headers=headers, data=data) as response:
            result = await response.json()
            return result.get('references', [])

async def comprehensive_search(keyword, page=1):
    async with aiohttp.ClientSession() as session:
        semaphore = asyncio.Semaphore(5)  # Max 5 concurrent requests
        tasks = [fetch_search(session, semaphore, config, keyword, page) 
                 for config in SEARCH_CONFIGS]
        results = await asyncio.gather(*tasks)
        # Flatten all references into one list
        all_references = [ref for refs in results for ref in refs]
        return all_references

asyncio.run(comprehensive_search("人工智能"))

Parameter Constraints

  • search_source: One of baidu_search, google_search, baidu_search_ai
  • mode: One of network, warehouse
  • When mode=warehouse, search is performed against the Elasticsearch index and ignores search_source
  • When mode=network, use search_source to select Baidu, Google, or Baidu AI search

Response Format

{
  "code": 200,
  "message": "success",
  "references": [
    {
      "title": "Article title",
      "sourceAddress": "https://example.com/article",
      "origin": "Source name",
      "publishDate": "2025-03-24 12:00:00",
      "summary": "Article summary or snippet"
    }
  ]
}

Usage Examples

Example 1: Search Baidu news

POST https://clb.ciglobal.cn/web_search
Headers: X-API-Key: <your_api_key>, Content-Type: application/x-www-form-urlencoded
Body (form): keyword=人工智能&search_source=baidu_search&mode=network&page=1

Example 2: Search Google news

POST https://clb.ciglobal.cn/web_search
Headers: X-API-Key: <your_api_key>, Content-Type: application/x-www-form-urlencoded
Body (form): keyword=AI&search_source=google_search&mode=network&page=1

Example 3: Search warehouse (ES index)

POST https://clb.ciglobal.cn/web_search
Headers: X-API-Key: <your_api_key>, Content-Type: application/x-www-form-urlencoded
Body (form): keyword=机器学习&mode=warehouse&page=1

Example 4: cURL

curl -X POST "https://clb.ciglobal.cn/web_search" \
  -H "X-API-Key: <your_api_key>" \
  -H "Content-Type: application/x-www-form-urlencoded" \
  -d "keyword=科技新闻&search_source=baidu_search&mode=network&page=1"

Error Codes

CodeMessageCause
401X-API-Key参数缺失Missing API key header
401ApiKey无效Invalid API key
400search_source参数错误Invalid search_source value
400mode参数错误Invalid mode value
400page参数错误Invalid page (non-integer or 0)

Integration Steps

  1. Verify Provider: Before publishing or using this skill, verify clb.ciglobal.cn’s trust model, privacy policy, data retention practices, and ownership information
  2. Obtain API Key: Visit https://clb.ciglobal.cn/apiKey/login to get your API key
  3. Configure Credentials: Set the GLOBAL_SEARCH_API_KEY environment variable with your API key

- Linux/Mac: export GLOBAL_SEARCH_API_KEY="your_api_key_here" - Windows: $env:GLOBAL_SEARCH_API_KEY="your_api_key_here"

  1. API Base URL: https://clb.ciglobal.cn
  2. Comprehensive Search: Only perform comprehensive search across all sources when the user explicitly requests broad multi-source coverage
  3. Sensitive Query Check: If the user’s query includes sensitive, personal, confidential, or internal data, refuse to transmit it and ask for a redacted or generalized query instead
  4. Make Request: Call POST /web_search with form-encoded parameters and include X-API-Key header
  5. Parse Response: Parse references from the response and use title, sourceAddress, summary as needed

平台分布

OpenClaw

95.55%
按下载量换算3,547

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

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