Token导航 LogoToken导航TokenDH.com
研究检索敏感数据clawhub未标认证来源可访问clear审计提醒

hub-client中心客户端

Agent Skill

hub-client 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。

总安装

1,996

周安装

84

GitHub Stars

公开资料未说明

下载量

699
OpenClaw

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

MIT-0

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:hub-client(中心客户端)
来源仓库:https://github.com/tangboheng/hub-client
安装命令:
openclaw skills install hub-client
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 hub-client

简介

将数据发布为服务并支持中心调用的集成工具。

  • 适用于企业内部服务市场搭建与管理。
  • 简化数据接口暴露与使用流程。适用宿主包括 OpenClaw,接入前应确认版本、权限和运行环境要求。
  • 需配置认证与权限控制机制。hub-client 属于研究检索类 Skill,可作为该场景下的辅助能力补充。
  • 建议查阅原始文档了解部署架构与安全策略。

SKILL.md

name
claw-service-hub
description
Service marketplace: publish data as services, consume hub services
version
1.0.0
homepage
https://github.com/TangBoheng/Claw-Service-Hub
metadata
openclaw
emoji
🔌
requires
bins
["python", "pip"]
env
["HUB_WS_URL"]
pip
["websockets", "aiohttp"]
primaryEnv
HUB_WS_URL
install
kind
pip
package
claw-service-hub
label
Install via pip

triggers: provider: - provide.*service - publish.*service - expose.*service - create.*service - make.*service - implement.*service consumer: - what services - list services - call.*service - use.*service - query.*data - fetch.*data

Tool Service Hub Skill

Overview

Enables subagents to:

  1. Provider Mode: Publish local data/capabilities as services for other subagents to call
  2. Consumer Mode: Discover and call services on the Hub

1. Publishing Services as a Provider

Complete Code Template

import asyncio
import os
import sys
from pathlib import Path

# === 1. Setup path ===
WORKSPACE_DIR = os.getenv('WORKSPACE_DIR', '/home/t/.openclaw/workspace-subagentX')
sys.path.insert(0, WORKSPACE_DIR)

from client.client import LocalServiceRunner

# === 2. Define your service capability ===

async def your_method(**params):
    """
    Service method
    params: Parameters passed by the caller
    Must return a dict
    """
    # Your business logic here
    result = {"status": "ok", "data": "..."}
    return result

# === 3. Start the service ===

async def main():
    runner = LocalServiceRunner(
        name="your-service-name",      # Service name (English, no spaces)
        description="Service description",  # English description
        hub_url=os.getenv("HUB_WS_URL", "ws://localhost:8765")
    )
    
    # Register methods (can register multiple)
    runner.register_handler("your_method", your_method)
    
    print(f"🚀 Starting service...")
    await runner.run()

if __name__ == "__main__":
    asyncio.run(main())

2. Calling Services as a Consumer

Complete Code Template

import asyncio
import os
import sys

WORKSPACE_DIR = os.getenv('WORKSPACE_DIR', '/home/t/.openclaw/workspace-subagentX')
sys.path.insert(0, WORKSPACE_DIR)

from client.skill_client import SkillQueryClient

async def main():
    # 1. Connect to Hub
    client = SkillQueryClient(
        hub_url=os.getenv("HUB_WS_URL", "ws://localhost:8765")
    )
    await client.connect()
    
    # 2. Discover services
    services = await client.discover()
    print(f"Discovered {len(services)} services")
    
    # 3. Find target service (filter by name)
    target = None
    target_name = "weather-service"  # Replace with your target service name
    for s in services:
        if target_name in s.get("name", ""):
            target = s
            break
    
    if not target:
        print(f"Service not found: {target_name}")
        return
    
    skill_id = target.get("skill_id")
    print(f"Using service: {target.get('name')}, skill_id: {skill_id}")
    
    # 4. Call the service
    result = await client.call_service(
        service_id=skill_id,
        method="your_method",      # Method name
        params={"key": "value"}    # Parameters
    )
    
    print(f"Result: {result}")
    
    await client.disconnect()

if __name__ == "__main__":
    asyncio.run(main())

3. Common Data Source Examples

3.1 File Data Service

from pathlib import Path

DATA_DIR = Path("/path/to/data")  # Change to actual directory

async def list_files(**params):
    ext = params.get("extension", "")
    pattern = f"*{ext}" if ext else "*"
    files = [f.name for f in DATA_DIR.glob(pattern) if f.is_file()]
    return {"files": files[:50], "total": len(files)}

async def read_file(**params):
    filename = params.get("filename")
    if not filename:
        return {"error": "filename is required"}
    
    filepath = DATA_DIR / filename
    if not filepath.exists():
        return {"error": f"File not found: {filename}"}
    
    # Read text files directly
    if filepath.suffix == '.txt':
        return {"content": filepath.read_text()[:1000]}
    
    # Return info for other files
    return {"filename": filename, "size": filepath.stat().st_size}

3.2 API Data Service

import aiohttp

async def fetch_data(**params):
    url = params.get("url")
    if not url:
        return {"error": "url is required"}
    
    try:
        async with aiohttp.ClientSession() as session:
            async with session.get(url, timeout=aiohttp.ClientTimeout(10)) as resp:
                data = await resp.json()
        return {"status": resp.status, "data": data}
    except Exception as e:
        return {"error": str(e)}

3.3 Weather Service (wttr.in)

import aiohttp

async def get_weather(**params):
    city = params.get("city", "Shanghai")
    url = f"https://wttr.in/{city}?format=j1"
    
    try:
        async with aiohttp.ClientSession() as session:
            async with session.get(url, timeout=aiohttp.ClientTimeout(10)) as resp:
                data = await resp.json()
        
        # Note: wttr.in returns structure at data.current_condition
        c = data.get("data", {}).get("current_condition", [{}])[0]
        return {
            "city": city,
            "temp": int(c.get("temp_C") or 0),
            "condition": c.get("weatherDesc", [{}])[0].get("value", "Unknown"),
            "humidity": c.get("humidity")
        }
    except Exception as e:
        return {"error": str(e)}

4. Workflow Examples

Combining Multiple Services

async def workflow():
    """Complete workflow combining multiple services"""
    client = SkillQueryClient("ws://localhost:8765")
    await client.connect()
    
    services = await client.discover()
    
    # Find required services
    weather = next((s for s in services if "weather" in s.get("name", "")), None)
    images = next((s for s in services if "image" in s.get("name", "")), None)
    
    results = {}
    
    # Call weather service
    if weather:
        w = await client.call_service(weather.get("skill_id"), "get_weather", {"city": "Shanghai"})
        results["weather"] = w.get("result", {})
    
    # Call image service
    if images:
        i = await client.call_service(images.get("skill_id"), "list_images", {"limit": 10})
        results["images"] = i.get("result", {})
    
    await client.disconnect()
    return results

5. Environment Configuration

Install Dependencies

pip install websockets aiohttp

Environment Variables

VariableDefaultDescription
HUB_WS_URLws://localhost:8765Hub WebSocket address
WORKSPACE_DIR/home/t/.../workspace-subagentXWorking directory

Start Hub Server (Optional)

cd Claw-Service-Hub
python -m server.main

# WebSocket: ws://0.0.0.0:8765
# REST API: http://0.0.0.0:3765

6. Troubleshooting

Issue 1: ImportError

Error: ModuleNotFoundError: No module named 'client'

Fix: Set sys.path correctly

import os
import sys
WORKSPACE_DIR = os.getenv('WORKSPACE_DIR', '/home/t/.openclaw/workspace-subagentX')
sys.path.insert(0, WORKSPACE_DIR)
from client.client import LocalServiceRunner

Issue 2: Service Registered but Cannot Be Called

Check:

  1. Is the Provider process still running?
  2. Is the method name correct (case-sensitive)?
  3. Is the parameter format correct?

Issue 3: API Returns None

Common Cause: Data structure parsing error

Fix: Print raw data first to confirm structure

async def get_data(**params):
    url = params.get("url")
    async with aiohttp.ClientSession() as session:
        async with session.get(url) as resp:
            data = await resp.json()
    print(f"Raw data: {data}")  # Add this debug line
    # Then parse based on actual structure
    return {"data": data}

Issue 4: Service Not Found

Fix: List all services first

services = await client.discover()
for s in services:
    print(f"{s.get('name')}: {s.get('skill_id')}")

Issue 5: Return Value Must Be Dict

Error: TypeError: ... got an unexpected keyword argument

Fix: Handlers must return dict

async def wrong():  # Wrong
    return "string"

async def right(**params):  # Correct
    return {"result": "value"}

Issue 6: Return Value Wrapped in 'result'

现象: After calling service, returns {'result': {'actual_data': '...'}}

说明: Hub wraps the handler's returned dict in the 'result' field

Fix: Use the return value directly, or extract as needed

result = await client.call_service(service_id, "method", params)
# result = {'result': {'temp': 25, 'city': 'Beijing'}}

# Method 1: Use directly (recommended)
data = result  # Already unpacked data

# Method 2: If explicit extraction needed
if 'result' in result:
    data = result['result']

7. Minimal Examples

Provider (5 lines)

import asyncio, os, sys
sys.path.insert(0, os.getenv('WORKSPACE_DIR','.'))
from client.client import LocalServiceRunner

async def hello(**p): return {"msg":"Hello!"}
r = LocalServiceRunner("demo","Demo Service",os.getenv("HUB_WS_URL","ws://localhost:8765"))
r.register_handler("hello", hello)
asyncio.run(r.run())

Consumer (6 lines)

import asyncio, os, sys
sys.path.insert(0, os.getenv('WORKSPACE_DIR','.'))
from client.skill_client import SkillQueryClient

async def main():
    c = SkillQueryClient()
    await c.connect()
    print([s.get("name") for s in await c.discover()])
    await c.disconnect()
asyncio.run(main())

8. File Structure

Claw-Service-Hub/
├── client/
│   ├── client.py           # LocalServiceRunner, ToolServiceClient
│   ├── skill_client.py     # SkillQueryClient
│   └── management_client.py
├── skills/
│   └── hub-client/
│       └── SKILL.md        # This file
└── server/
    └── main.py             # Hub Server

9. Publishing Flow

  1. Ensure Hub Server is running (ws://localhost:8765)
  2. Provider: Run python your_service.py to register service
  3. Consumer: Connect to Hub to discover services
  4. Consumer: Call service to get results

10. Key Authorization Mechanism (Optional)

10.1 Overview

Optional key authorization mechanism for controlling service access.

Features:

  • Time dimension: Key validity period (customizable by Provider)
  • Count dimension: Maximum call count (customizable by Provider)
  • Dual verification: Provider self-management + Hub verification

10.2 Provider Side - Set Lifecycle Policy

from client.client import LocalServiceRunner

# Create service
runner = LocalServiceRunner(
    name="my-protected-service",
    description="Service requiring authorization",
    hub_url=os.getenv("HUB_WS_URL", "ws://localhost:8765")
)

# Set default lifecycle policy
runner.set_lifecycle_policy(
    duration_seconds=3600,  # Default 1 hour validity
    max_calls=100           # Default 100 calls
)

# Set custom policy (optional)
runner.set_custom_policy(
    condition="premium",     # Policy name
    duration_seconds=86400,  # 24 hours
    max_calls=1000           # 1000 calls
)

# Register method
async def get_data(**params):
    return {"data": "secret data"}

runner.register_handler("get_data", get_data)

# Start service
print(f"🚀 Starting authorized service...")
await runner.run()

10.3 Consumer Side - Request Key and Call

from client.skill_client import SkillQueryClient

async def main():
    client = SkillQueryClient(
        hub_url=os.getenv("HUB_WS_URL", "ws://localhost:8765")
    )
    await client.connect()
    
    # Discover services
    services = await client.discover()
    target = next((s for s in services if "protected" in s.get("name", "")), None)
    
    if not target:
        print("Service not found")
        return
    
    service_id = target.get("skill_id")
    
    # Method 1: Direct call (if service doesn't require Key)
    # result = await client.call_service(service_id, "get_data", {})
    
    # Method 2: Request Key first, then call (recommended)
    key_info = await client.request_key(
        service_id=service_id,
        purpose="Daily data query"
    )
    
    if key_info.get("success"):
        key = key_info["key"]
        lifecycle = key_info["lifecycle"]
        
        print(f"✅ Key obtained successfully")
        print(f"   Key: {key[:20]}...")
        print(f"   Validity: {lifecycle.get('remaining_time')} seconds")
        print(f"   Remaining calls: {lifecycle.get('remaining_calls')} times")
        
        # Call service with Key
        result = await client.call_service(
            service_id=service_id,
            method="get_data",
            params={},
            key=key  # Carry the Key
        )
        
        print(f"📥 Result: {result}")
    else:
        print(f"❌ Key request failed: {key_info.get('reason')}")
    
    await client.disconnect()

asyncio.run(main())

10.4 Handling Key Verification Failure

async def call_with_fallback(client, service_id, method, params):
    """Call with automatic retry"""
    
    # Try without Key first
    result = await client.call_service(service_id, method, params)
    
    # Check if Key is required
    if result.get("error") and "Key" in result.get("error", ""):
        print("Key required, requesting authorization...")
        
        key_info = await client.request_key(service_id, "Auto request")
        if key_info.get("success"):
            key = key_info["key"]
            # Retry with Key
            result = await client.call_service(
                service_id, method, params, key=key
            )
    
    return result

10.5 Message Protocol

Message TypeDirectionDescription
lifecycle_policyProvider→HubRegister lifecycle policy
key_requestConsumer→Hub→ProviderRequest Key
key_responseProvider→Hub→ConsumerReturn Key (approve/reject)
key_revokeProvider→HubRevoke Key
call_service (with key)Consumer→HubCall service with Key

10.6 Lifecycle Parameters

# Provider registers policy
{
    "duration_seconds": 3600,    # Validity duration (seconds)
    "max_calls": 100,             # Maximum call count
    "custom_policies": {         # Optional: custom policies
        "premium": {
            "duration_seconds": 86400,
            "max_calls": 1000
        }
    }
}

# Key verification result
{
    "valid": True,
    "key": "key_abc123...",
    "lifecycle": {
        "expires_at": "2026-03-20T03:47:00Z",
        "max_calls": 100,
        "call_count": 5,
        "remaining_calls": 95,
        "remaining_time": 3200
    }
}

10.7 Subagent Usage Tips

For Provider (Service Publisher):

  1. If service requires authorization, set reasonable limits in set_lifecycle_policy
  2. Recommended defaults: max_calls=100, duration_seconds=3600
  3. Can set different policies for different consumers

For Consumer (Service Caller):

  1. Try direct call first, request Key if it fails
  2. Save Key to avoid repeated requests
  3. Check lifecycle.get('remaining_calls') to avoid running out of calls
  4. Watch lifecycle.get('remaining_time') and renew as needed

License

MIT License

Copyright (c) 2026 OpenClaw

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

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

补充不同宿主或平台的使用分布数据

能力 5

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

OpenClaw

84.27%
按下载量换算589

安全审计

VirusTotal

未展示

ClawScan

可疑

Static analysis

通过

权限和风险

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。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