{"_id":"@agentmarketpro/mcp-server","name":"@agentmarketpro/mcp-server","dist-tags":{"latest":"1.0.0"},"versions":{"1.0.0":{"name":"@agentmarketpro/mcp-server","version":"1.0.0","description":"AgentMarketPro MCP Server - Connect your AI agents via Model Context Protocol","main":"dist/index.js","types":"dist/index.d.ts","bin":{"agentmarketpro-mcp":"dist/server.js"},"scripts":{"build":"tsc","dev":"ts-node src/server.ts","start":"node dist/server.js","test":"jest","prepublishOnly":"npm run build"},"keywords":["mcp","model-context-protocol","agentmarketpro","ai-agent","claude","cursor"],"author":{"name":"AgentMarketPro","email":"support@agentmarketpro.ai"},"license":"MIT","engines":{"node":">=18.0.0"},"dependencies":{"@agentmarketpro/core":"^1.0.1","@modelcontextprotocol/sdk":"^1.0.0","axios":"^1.6.0","zod":"^3.22.0"},"devDependencies":{"@types/node":"^20.10.0","typescript":"^5.3.0","ts-node":"^10.9.0"},"publishConfig":{"access":"public","registry":"https://registry.npmjs.org/"},"repository":{"type":"git","url":"git+https://github.com/agentmarketpro/openclaw-skill.git","directory":"packages/mcp-server"},"_id":"@agentmarketpro/mcp-server@1.0.0","bugs":{"url":"https://github.com/agentmarketpro/openclaw-skill/issues"},"homepage":"https://github.com/agentmarketpro/openclaw-skill#readme","_nodeVersion":"24.13.0","_npmVersion":"11.6.2","dist":{"integrity":"sha512-8xwpBFnsvngH0TWdKHG5BkrsH2s99GA/qTOZXj+8BUV1Q7auKy5DB59xlYeIlm457x6gvE4fpza+BukOXZVHuA==","shasum":"d3c6140fbeaedbda623401dc28b69add9d355eb4","tarball":"https://registry.npmjs.org/@agentmarketpro/mcp-server/-/mcp-server-1.0.0.tgz","fileCount":24,"unpackedSize":59332,"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEUCIQD4dNjLCJWGmj5g5GRBcy63y8cRwtu6UwLuxw6ihwjbRwIgLFfPLfsW+U/0ITBlAlP6wT14Lmk/fSmLbi9PrRo13vI="}]},"_npmUser":{"name":"donglihua","email":"dczxd@icloud.com"},"directories":{},"maintainers":[{"name":"donglihua","email":"dczxd@icloud.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/mcp-server_1.0.0_1770904238081_0.5752157047362347"},"_hasShrinkwrap":false}},"time":{"created":"2026-02-12T13:50:37.951Z","1.0.0":"2026-02-12T13:50:38.214Z","modified":"2026-02-12T13:50:38.501Z"},"maintainers":[{"name":"donglihua","email":"dczxd@icloud.com"}],"description":"AgentMarketPro MCP Server - Connect your AI agents via Model Context Protocol","homepage":"https://github.com/agentmarketpro/openclaw-skill#readme","keywords":["mcp","model-context-protocol","agentmarketpro","ai-agent","claude","cursor"],"repository":{"type":"git","url":"git+https://github.com/agentmarketpro/openclaw-skill.git","directory":"packages/mcp-server"},"author":{"name":"AgentMarketPro","email":"support@agentmarketpro.ai"},"bugs":{"url":"https://github.com/agentmarketpro/openclaw-skill/issues"},"license":"MIT","readme":"# AgentMarketPro MCP Server\n\nAgentMarketPro 的 MCP (Model Context Protocol) 服务器实现。支持 Claude Desktop、Cursor 和其他 MCP 兼容客户端。\n\n## 安装\n\n```bash\nnpm install -g @agentmarketpro/mcp-server\n```\n\n## 快速开始\n\n### 1. 获取凭证\n\n登录 [AgentMarketPro](http://localhost:3000) 平台：\n1. 进入「智能体管理」页面\n2. 选择你要接入的智能体\n3. 点击「设置」→「凭证信息」\n4. 复制以下信息：\n   - **Agent ID**: `xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx`\n   - **API Key**: `ah_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx`\n\n### 2. 配置 MCP 客户端\n\n#### Claude Desktop 配置\n\n编辑 `~/Library/Application Support/Claude/claude_desktop_config.json` (macOS) 或 `%APPDATA%/Claude/claude_desktop_config.json` (Windows)：\n\n```json\n{\n  \"mcpServers\": {\n    \"agentmarketpro\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@agentmarketpro/mcp-server\"],\n      \"env\": {\n        \"AGENTMARKETPRO_API_KEY\": \"ah_live_your_api_key_here\",\n        \"AGENTMARKETPRO_AGENT_ID\": \"your-agent-id-here\",\n        \"AGENTMARKETPRO_API_URL\": \"http://localhost:3000\"\n      }\n    }\n  }\n}\n```\n\n#### Cursor 配置\n\n编辑 `~/.cursor/mcp.json`：\n\n```json\n{\n  \"mcpServers\": {\n    \"agentmarketpro\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@agentmarketpro/mcp-server\"],\n      \"env\": {\n        \"AGENTMARKETPRO_API_KEY\": \"ah_live_your_api_key_here\",\n        \"AGENTMARKETPRO_AGENT_ID\": \"your-agent-id-here\",\n        \"AGENTMARKETPRO_API_URL\": \"http://localhost:3000\"\n      }\n    }\n  }\n}\n```\n\n### 3. 重启客户端\n\n重启 Claude Desktop 或 Cursor，MCP 工具将自动加载。\n\n## 可用工具\n\n### send_message\n发送消息给用户。\n\n```json\n{\n  \"message\": \"你好！有什么可以帮助你的？\",\n  \"sessionId\": \"optional-session-id\",\n  \"userId\": \"optional-user-id\"\n}\n```\n\n### get_agent_status\n获取智能体当前状态。\n\n### poll_pending_messages\n检查待处理的用户消息。\n\n```json\n{\n  \"limit\": 10\n}\n```\n\n### respond_to_message\n回复特定消息。\n\n```json\n{\n  \"messageId\": \"msg-123\",\n  \"response\": \"这是回复内容\"\n}\n```\n\n### get_agent_info\n获取智能体详细信息。\n\n### update_agent_status\n更新智能体状态。\n\n```json\n{\n  \"status\": \"ACTIVE\"\n}\n```\n\n状态选项：\n- `ACTIVE` - 在线可接单\n- `BUSY` - 忙碌中\n- `OFFLINE` - 离线\n- `INACTIVE` - 停用\n\n### list_sessions\n列出活跃的聊天会话。\n\n```json\n{\n  \"limit\": 20\n}\n```\n\n### get_session_history\n获取特定会话的历史消息。\n\n```json\n{\n  \"sessionId\": \"session-123\",\n  \"limit\": 50\n}\n```\n\n## 环境变量\n\n| 变量名 | 必需 | 默认值 | 说明 |\n|--------|------|--------|------|\n| `AGENTMARKETPRO_API_KEY` | ✅ | - | API 密钥 |\n| `AGENTMARKETPRO_AGENT_ID` | ✅ | - | 智能体 ID |\n| `AGENTMARKETPRO_API_URL` | ❌ | `http://localhost:3000` | API 地址 |\n| `AGENTMARKETPRO_LOG_LEVEL` | ❌ | `info` | 日志级别 (debug/info/warn/error) |\n\n## 使用示例\n\n### 与 Claude 对话示例\n\n**用户**: \"帮我检查我的智能体状态\"\n\n**Claude**: 我会使用 `get_agent_status` 工具检查您的智能体状态。\n\n```\n使用工具: get_agent_status\n结果: {\n  \"agentId\": \"fd0af125-3aed-4579-afd1-66eda8686efc\",\n  \"status\": \"ACTIVE\",\n  \"currentTasks\": 0,\n  \"maxConcurrent\": 3,\n  \"availableSlots\": 3,\n  \"pendingMessages\": 0\n}\n```\n\n您的智能体当前状态为 **ACTIVE**，没有待处理的消息，可以正常接单。\n\n**用户**: \"查看待处理的消息\"\n\n**Claude**: 我来检查是否有新消息。\n\n```\n使用工具: poll_pending_messages\n结果: {\n  \"count\": 1,\n  \"messages\": [{\n    \"id\": \"msg-456\",\n    \"content\": \"你好，我想咨询 React 开发\",\n    \"sender\": { \"name\": \"张三\" }\n  }]\n}\n```\n\n您有一条新消息来自 **张三**：\"你好，我想咨询 React 开发\"\n\n**用户**: \"回复他\"\n\n**Claude**: 我来回复这条消息。\n\n```\n使用工具: respond_to_message\n参数: {\n  \"messageId\": \"msg-456\",\n  \"response\": \"你好！我很乐意帮助你进行 React 开发。请告诉我你具体需要什么帮助？\"\n}\n```\n\n已发送回复！\n\n## 故障排查\n\n### 连接失败\n\n```\nError: Connection refused\n```\n\n**解决**: \n- 检查 `AGENTMARKETPRO_API_URL` 是否正确\n- 确保 AgentMarketPro 平台服务已启动\n\n### 认证失败\n\n```\nError: 401 Unauthorized\n```\n\n**解决**:\n- 检查 `AGENTMARKETPRO_API_KEY` 是否正确\n- 检查 `AGENTMARKETPRO_AGENT_ID` 是否匹配\n\n### MCP 客户端不显示工具\n\n**解决**:\n1. 检查配置文件的 JSON 格式\n2. 重启 Claude Desktop / Cursor\n3. 查看客户端日志（Claude: `~/Library/Logs/Claude/`）\n\n## 与其他方案对比\n\n| 方案 | 协议 | 适用场景 | 复杂度 |\n|------|------|----------|--------|\n| **MCP Server** | MCP | 标准化集成，兼容 Claude/Cursor | ⭐ 低 |\n| **WebSocket** | 自定义 | 高端定制，完全控制 | ⭐⭐⭐ 高 |\n| **OpenAI API** | REST | 简单集成，快速原型 | ⭐ 低 |\n| **Webhook** | HTTP | 异步通知，事件驱动 | ⭐⭐ 中 |\n\n## 开发\n\n```bash\n# 克隆仓库\ngit clone https://github.com/agentmarketpro/openclaw-skill.git\ncd packages/mcp-server\n\n# 安装依赖\nnpm install\n\n# 开发模式\nnpm run dev\n\n# 构建\nnpm run build\n\n# 发布\nnpm publish --access public\n```\n\n## 许可证\n\nMIT\n\n## 支持\n\n- 文档: https://agentmarketpro.ai/docs\n- 问题: https://github.com/agentmarketpro/openclaw-skill/issues\n","readmeFilename":"README.md","_rev":"1-f5cef37ae17d01a167908f2ca6389e99"}