1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16

# 使用 npm 安装
npm install -g openclaw@latest
# 或使用 pnpm
pnpm add -g openclaw@latest

# 启动安装向导
openclaw onboard --install-daemon

# 运行 Gateway
openclaw gateway --port 18789 --verbose

# 发送消息测试
openclaw message send --to +1234567890 --message "Hello from OpenClaw"

# 与助手对话
openclaw agent --message "Ship checklist" --thinking high

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

// OpenClaw 配置文件示例
{
  "gateway": {
    "port": 18789,
    "verbose": true
  },
  "channels": {
    "discord": {
      "token": "YOUR_DISCORD_BOT_TOKEN",
      "dmPolicy": "pairing"
    },
    "whatsapp": {
      "phoneId": "YOUR_PHONE_ID",
      "token": "YOUR_WHATSAPP_TOKEN"
    }
  },
  "models": {
    "primary": {
      "provider": "openai",
      "model": "gpt-4o"
    }
  }
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14

# 从源码编译安装
git clone https://github.com/ultraworkers/claw-code.git
cd claw-code
npm install
npm run build

# 或使用 npm 全局安装
npm install -g claw-code

# 初始化配置
claw-code init

# 开始使用
claw-code --interactive

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14

# 使用 claw-code 生成的示例代码
def fibonacci(n, memo={}):
    """计算第 n 个斐波那契数"""
    if n in memo:
        return memo[n]
    if n <= 2:
        return 1
    memo[n] = fibonacci(n-1, memo) + fibonacci(n-2, memo)
    return memo[n]

# 测试
if __name__ == "__main__":
    for i in range(1, 11):
        print(f"fib({i}) = {fibonacci(i)}")

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12

# 全局安装
npm install -g opencode

# 初始化项目
opencode init my-project
cd my-project

# 运行代理
opencode run --task "Create a REST API for user management"

# 交互模式
opencode chat

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18

// opencode 自动生成的 Express API 示例
import express, { Request, Response } from 'express';
import { UserController } from './controllers/UserController';

const app = express();
app.use(express.json());

// 用户相关路由
app.get('/api/users', UserController.list);
app.get('/api/users/:id', UserController.get);
app.post('/api/users', UserController.create);
app.put('/api/users/:id', UserController.update);
app.delete('/api/users/:id', UserController.delete);

const PORT = process.env.PORT || 3000;
app.listen(PORT, () => {
  console.log(`Server running on port ${PORT}`);
});

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13

# 使用 pip 安装
pip install hermes-agent

# 或从源码安装
git clone https://github.com/NousResearch/hermes-agent.git
cd hermes-agent
pip install -e .

# 初始化
hermes init

# 启动代理
hermes run --mode interactive

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20

from hermes import Agent, MemorySystem

# 创建具有记忆系统的代理
agent = Agent(
    name="ResearchAssistant",
    memory=MemorySystem(
        long_term_storage="./knowledge_base",
        short_term_limit=100
    ),
    learning_rate=0.01
)

# 使用代理进行任务
async def research_topic(topic: str):
    await agent.learn(topic)
    return await agent.analyze(topic)

# 与代理对话
response = await agent.chat("Explain machine learning basics")
print(response)

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14

# 从源码编译
git clone https://github.com/openai/codex.git
cd codex
cargo build --release

# 或使用 cargo install
cargo install codex-cli

# 基本使用
codex "Write a Python script to process CSV files"
codex --interactive

# 在 CI 中使用
codex --input task.txt --output solution.py

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19

// codex 的配置文件示例
use codex::Config;

fn main() {
    let config = Config::builder()
        .model("claude-code")
        .temperature(0.7)
        .max_tokens(2048)
        .sandbox(true)
        .build();

    let agent = codex::Agent::new(config);
    
    let response = agent
        .execute("Create a function to parse JSON efficiently")
        .unwrap();
    
    println!("{}", response);
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13

# 安装 deer-flow
git clone https://github.com/bytedance/deer-flow.git
cd deer-flow
pip install -e .

# 初始化配置
deer-flow init

# 运行长时间任务
deer-flow run --task "Analyze this codebase and generate documentation" --duration hours

# API 模式
deer-flow serve --port 8080

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

from deer_flow import Agent, SandBox, Memory, Tools

# 创建带有沙箱和记忆的代理
agent = Agent(
    name="SuperResearcher",
    sandbox=SandBox(
        timeout=3600,  # 1小时超时
        memory_limit="4GB"
    ),
    memory=Memory(
        storage="./research_memory",
        max_items=10000
    ),
    tools=Tools.all(),  # 或选择特定工具
    subagents=["researcher", "coder", "tester"]
)

# 处理复杂任务
async def comprehensive_research(query: str):
    result = await agent.run(
        task=query,
        max_duration=7200,  # 最长2小时
        checkpoints=[0.25, 0.5, 0.75]  # 进度检查点
    )
    return result

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14

# 安装 n8n
npm install n8n -g

# 启动 n8n
n8n start

# 或使用 Docker
docker run -it --rm \
  -p 5678:5678 \
  -v ~/.n8n:/home/node/.n8n \
  n8nio/n8n

# 导入工作流
n8n import:workflow --input workflow.json

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

{
  "name": "每日开源项目通知",
  "nodes": [
    {
      "name": "GitHub Trigger",
      "type": "n8n-nodes-base.gitHubTrigger",
      "parameters": {
        "event": "star",
        "repositories": ["openclaw/openclaw"]
      }
    },
    {
      "name": "Filter",
      "type": "n8n-nodes-base.filter",
      "parameters": {
        "conditions": {
          "value": {
            ["stars"]: [">", 10000]
          }
        }
      }
    },
    {
      "name": "Send to Discord",
      "type": "n8n-nodes-base.discord",
      "parameters": {
        "webhookUri": "YOUR_WEBHOOK_URL",
        "text": "={{ '⭐ New popular repo: ' + $json.name }}"
      }
    }
  ]
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

# 克隆仓库
git clone https://github.com/shareAI-lab/learn-claude-code.git
cd learn-claude-code

# 直接运行
chmod +x run.sh
./run.sh

# 或自定义配置
cat config.yaml

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15

#!/bin/bash
# learn-claude-code 核心脚本示例

# 代理主循环
while true; do
    echo -n "You: "
    read -r input
    
    # 调用模型
    response=$(curl -s -X POST "http://localhost:11434/api/generate" \
        -H "Content-Type: application/json" \
        -d "{\"model\": \"llama3\", \"prompt\": \"$input\"}" | jq -r '.response')
    
    echo -e "\nAssistant: $response\n"
done

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13

# 使用 npm 安装
npm install -g oh-my-openagent

# 或从源码
git clone https://github.com/code-yeongyu/oh-my-openagent.git
cd oh-my-openagent
npm install
npm run build

# 创建新项目
omo create my-agent
cd my-agent
npm run dev

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45

import { Agent, Workflow, Tools } from 'oh-my-openagent';

// 定义智能体
const codingAgent = new Agent({
  name: 'SeniorDeveloper',
  model: 'claude-opus-4',
  tools: [
    Tools.fileSystem,
    Tools.git,
    Tools.docker,
    Tools.terminal,
    Tools.search
  ],
  systemPrompt: `You are a senior software engineer.
You always write clean, well-documented code.
You prefer TypeScript and Rust.`
});

// 定义工作流
const codingWorkflow = new Workflow({
  name: 'FeatureDevelopment',
  steps: [
    {
      name: 'Analyze Requirements',
      agent: codingAgent,
      task: 'Analyze the feature requirements and create a technical design'
    },
    {
      name: 'Implement',
      agent: codingAgent,
      task: 'Implement the feature with tests'
    },
    {
      name: 'Review',
      agent: codingAgent,
      task: 'Review the implementation and suggest improvements'
    }
  ]
});

// 执行工作流
async function developFeature(feature: string) {
  const result = await codingWorkflow.execute({ feature });
  return result;
}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13

# 安装 shannon
npm install -g shannon-lite

# 或使用 Docker
docker pull keygraph/shannon-lite
docker run -it --rm keygraph/shannon-lite

# 基本使用
shannon scan --target https://api.example.com
shannon analyze --path ./src

# 生成报告
shannon report --format json --output security-report.json

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

import { Shannon, ScanConfig, VulnerabilityReport } from 'shannon-lite';

// 配置扫描参数
const config: ScanConfig = {
  targetUrl: 'https://api.example.com',
  scanType: 'comprehensive',
  auth: {
    type: 'bearer',
    token: process.env.API_TOKEN
  },
  checks: [
    'sql-injection',
    'xss',
    'csrf',
    'auth-bypass',
    'idor',
    'rate-limit'
  ],
  verbose: true,
  outputReport: './reports/security-report.json'
};

// 执行安全扫描
async function securityAudit(url: string): Promise<VulnerabilityReport> {
  const shannon = new Shannon(config);
  
  const results = await shannon.run();
  
  // 分析结果
  const criticalVulnerabilities = results.vulnerabilities
    .filter(v => v.severity === 'critical');
  
  if (criticalVulnerabilities.length > 0) {
    console.log('Critical vulnerabilities found!');
    console.log(criticalVulnerabilities);
  }
  
  return results;
}