经 AI Skill Hub 精选评估,人机交互MCP 获评「推荐使用」。这款MCP工具在功能完整性、社区活跃度和易用性方面表现出色,AI 评分 7.5 分,适合有一定技术背景的用户使用。
人机交互MCP 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
人机交互MCP 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
# 方式一:通过 Claude Code CLI 一键安装
claude skill install https://github.com/LIghtJUNction/humen-mcp
# 方式二:手动配置 claude_desktop_config.json
{
"mcpServers": {
"----mcp": {
"command": "npx",
"args": ["-y", "humen-mcp"]
}
}
}
# 配置文件位置
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%/Claude/claude_desktop_config.json
# 安装后在 Claude 对话中直接使用 # 示例: 用户: 请帮我用 人机交互MCP 执行以下任务... Claude: [自动调用 人机交互MCP MCP 工具处理请求] # 查看可用工具列表 # 在 Claude 中输入:"列出所有可用的 MCP 工具"
// claude_desktop_config.json 配置示例
{
"mcpServers": {
"____mcp": {
"command": "npx",
"args": ["-y", "humen-mcp"],
"env": {
// "API_KEY": "your-api-key-here"
}
}
}
}
// 保存后重启 Claude Desktop 生效
<p> A human-in-the-loop MCP server for moments where agents need judgment, visual review, account-bound action, or a short escalation path. </p>
<p> <a href="README.md">English</a> · <a href="README.zh-CN.md">简体中文</a> · <a href="#why-people-care">Why it matters</a> · <a href="#online-panel">Online panel</a> · <a href="#deploy-with-an-agent">Deploy</a> · <a href="#security-model">Security</a> · <a href="#technical-details">Technical details</a> </p> </div>
<p align="center"> <a href="https://github.com/LIghtJUNction/humen-mcp/actions/workflows/release.yml"><img alt="Release workflow" src="https://github.com/LIghtJUNction/humen-mcp/actions/workflows/release.yml/badge.svg" /></a> <a href="https://github.com/LIghtJUNction/humen-mcp/releases"><img alt="GitHub release" src="https://img.shields.io/github/v/release/LIghtJUNction/humen-mcp?display_name=tag&sort=semver" /></a> <a href="Cargo.toml"><img alt="MIT license" src="https://img.shields.io/badge/license-MIT-green.svg" /></a> <a href="Cargo.toml"><img alt="Rust 2024" src="https://img.shields.io/badge/Rust-2024-f46623?logo=rust&logoColor=white" /></a> <a href="https://modelcontextprotocol.io/"><img alt="MCP" src="https://img.shields.io/badge/MCP-human--in--the--loop-6f42c1" /></a> <a href="Cargo.toml"><img alt="Axum backend" src="https://img.shields.io/badge/backend-Axum-005f73" /></a> <a href="humen-mcp-webui/package.json"><img alt="React 19 frontend" src="https://img.shields.io/badge/frontend-React%2019-61dafb?logo=react&logoColor=111111" /></a> <a href="humen-mcp-webui/package.json"><img alt="Bun runtime" src="https://img.shields.io/badge/runtime-Bun-000000?logo=bun&logoColor=white" /></a> <a href="#deploy-with-an-agent"><img alt="AUR packages" src="https://img.shields.io/badge/packages-AUR-1793d1?logo=archlinux&logoColor=white" /></a> <a href="#security-model"><img alt="Passkeys and OAuth" src="https://img.shields.io/badge/auth-Passkeys%20%2B%20GitHub%20OAuth-2ea44f" /></a> </p>
humen-mcp lets an agent ask a logged-in person for help without leaving the MCP loop. The agent calls a tool, the server creates a task, the human answers in the web UI, and the agent receives the result.
Use it when a workflow needs a person for one small but important step: choosing between options, reading an image, checking a website, retrieving an account-bound code, approving a risky action, or escalating an ambiguous decision.
For Simplified Chinese documentation, see README.zh-CN.md.
Copy this command, run it anywhere with internet access, then send the fetched prompt to the agent that will deploy humen-mcp on your server:
curl -fsSL https://raw.githubusercontent.com/LIghtJUNction/humen-mcp/main/docs/AGENT_DEPLOY_PROMPT.md
The prompt tells the agent to inspect the server, install the package, configure the site root and /mcp, verify the MCP endpoint, and keep admin secrets out of public output.
<details> <summary><strong>Manual Server Deployment Guide</strong></summary>
This is the short human-run path for an Arch Linux server. See docs/DEPLOYMENT.md for the longer deployment note.
```bash paru -S humen-mcp-git
Run paru through a login shell for the normal AUR user:
sudo -iu arch bash -lc 'unset GIT_DIR GIT_WORK_TREE; cd ~; paru -S humen-mcp-git'
Plugins are declarative JSON or TOML manifests loaded at startup from HUMEN_PLUGIN_DIR. A manifest can contribute request templates, route strategies, scoring rules, and third-party channel declarations. The server exposes loaded plugins through list_humen_plugins, and agents can create an async request from a template with create_humen_request_from_template.
Template ids use plugin-id/template-id. Template text supports simple {{variable}} substitution from the tool call's variables object. Plugin authors can use the published humen-mcp-sdk crate to build and validate manifests before distribution.
Example manifest: examples/community-release-plugin.json.
</details>
<details> <summary><strong>HTTP and WebSocket API</strong></summary>
Authenticated UI APIs:
GET /api/meGET /api/me/profilePOST /api/me/profileGET /api/passkeysPOST /api/passkeys/register/startPOST /api/passkeys/register/finishPOST /api/passkeys/{id}/deleteGET /api/agent/accessPOST /api/agent/secretGET /api/requestsPOST /api/requests/{id}/answerGET /api/sentGET /api/tasksPOST /api/tasks/{id}/statusGET /api/trashPOST /api/trash/clearGET /api/users/onlineGET /api/users/search?q=...POST /api/humans/ratePOST /api/humans/reportGET /api/leaderboardGET /api/tagsGET /api/friendsPOST /api/friendsPOST /api/friends/{email}/acceptPOST /api/friends/{email}/removeGET /api/wsAdmin APIs:
GET /api/admin/usersPOST /api/admin/usersPOST /api/admin/users/{email}POST /api/admin/users/{email}/kickGET /api/admin/reportsGET /api/admin/settingsPOST /api/admin/settingsGET /api/admin/updatePOST /api/admin/updateGET /api/admin/webhooksPOST /api/admin/webhooksWebSocket events:
{ "type": "request_created", "request": {} }
{ "type": "request_answered", "id": "...", "answer": {} }
{ "type": "request_expired", "id": "...", "expired_request": {} }
{ "type": "trash_cleaned", "removed_count": 1 }
{ "type": "presence_changed", "online_count": 1 }
</details>
<details> <summary><strong>Configuration</strong></summary>
env.example documents all supported variables. Important production values:
HUMEN_BIND=127.0.0.1:8787
HUMEN_PUBLIC_BASE_URL=https://your-domain.example
HUMEN_WEB_DIST=/usr/share/humen-mcp/web
HUMEN_USERS_FILE=/var/lib/humen-mcp/users.json
HUMEN_DB_FILE=/var/lib/humen-mcp/humen-mcp.sqlite3
HUMEN_ADMIN_EMAIL=<admin-email>
HUMEN_ADMIN_PASSWORD=<generated-admin-password>
HUMEN_SESSION_SECRET=<generated-session-secret>
HUMEN_TRASH_RETENTION_SECONDS=604800
HUMEN_CLEANUP_INTERVAL_SECONDS=60
HUMEN_SELF_UPDATE_COMMAND=/usr/bin/sudo -n /usr/bin/systemctl start humen-mcp-self-update.service
HUMEN_SELF_UPDATE_TIMEOUT_SECONDS=30
HUMEN_PLUGIN_DIR=/etc/humen-mcp/plugins
HUMEN_NODE_ID=root
HUMEN_FEDERATION_FILE=/etc/humen-mcp/federation.toml
Production note: after AUR install, make sure /etc/humen-mcp.env uses the packaged web directory:
HUMEN_WEB_DIST=/usr/share/humen-mcp/web
If this is left as ./humen-mcp-webui/dist, https://your-domain.example/ will return 404 because the service runs from /var/lib/humen-mcp.
</details>
<details> <summary><strong>Arch Deployment</strong></summary>
Install with an AUR helper as a normal user, not root:
```bash paru -S humen-mcp-git
高质量的开源MCP工具,值得关注
该工具未明确声明开源协议,商业使用前请联系原作者确认授权范围,避免侵权风险。
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
AI Skill Hub 点评:人机交互MCP 的核心功能完整,质量良好。对于Claude Desktop / Claude Code 用户来说,这是一个值得纳入个人工具库的选择。建议先在非生产环境试用,再逐步推广。
| 原始名称 | humen-mcp |
| 原始描述 | 开源MCP工具:Human Interaction MCP Tool。⭐8 · Rust |
| Topics | agentsmcprust |
| GitHub | https://github.com/LIghtJUNction/humen-mcp |
| 语言 | Rust |
收录时间:2026-06-12 · 更新时间:2026-06-12 · License:未公布 · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端