mcp-universal-gentes
Universal MCP Lab
用于验证在稳定 MCP 表面背后的动态能力获取的最小概念验证。
假设
一个智能体以六个稳定工具开始。它连接一个 OpenAPI 服务,发现新的内部能力,检查它们并调用它们,而无需修改已发布的 MCP 目录。
预期的稳定表面:
services.connectservices.disconnectservices.listcapabilities.listcapabilities.describecapabilities.invoke
PoC 不提供对设备的原始访问,不暴露任意 shell,也不存储凭据。
Related MCP server: openapi-mcp
本地运行测试
python -m unittest discover -s tests -v
python scripts/demo.py通过 stdio 实现真实 MCP
第 3 阶段包含一个具有六个稳定工具的真实 MCP 服务器。安装声明的依赖项并让 MCP 主机启动该进程:
pip install -e .
python -m entradas.mcp.server该进程不显示界面,也不应写入 stdout:它通过 stdin 等待 MCP 消息。测试 tests/test_mcp_protocol.py 充当真实 MCP 客户端,验证 tools/list,连接合成 OpenAPI,并确认公共目录未更改。
这种本地传输是刻意受限的:不发布 HTTP,不包含 OAuth,不连接 PC,也不启用写批准。
要在实验室中启用 HTTP 读取,请在启动前配置本地允许列表。通过 MCP 发送的 OpenAPI 文档无法修改该目标:
$env:UNIVERSAL_MCP_READ_BASE_URLS = '{"customers":"https://api.example.test"}'
python -m entradas.mcp.server仅对已配置的服务执行 GET;有超时、大小限制、强制 JSON、无重定向或重试。该套件不使用互联网:它启动一个本地临时端点来验证此流程。
所有者的本地节点
PC 节点是独立的本地 MCP 进程,而不是远程控制功能。只有设备所有者才能启动它并设置明确文件夹:
$env:UNIVERSAL_MCP_OWNER_ALLOWED_ROOT = 'C:\carpeta-autorizada'
python -m entradas.pc_node.server其唯一工具是 pc.read_text(relative_path)。它不包括 shell、写入、磁盘浏览、浏览器、进程或网络侦听器。网关与此节点之间的隧道和身份验证属于后续阶段。
写入的 Action Boundary
写入不接受 approved=true。第一个请求返回确认引用;受信任的人工界面必须在 MCP 之外批准该确切操作。生成的授权是一次性的,绑定到相同参数,并在 executor 之前消耗。这尚未在 PC 节点上启用写入。
状态
架构 + 实验室核心 + 深度 ECA 已准备好。请参见 STATUS.md 和 docs/PLAYBOOK.md。
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- FlicenseAqualityDmaintenanceExposes two MCP tools (discover and execute) that enable agents to query an OpenAPI schema via natural language and execute matched API operations.2
- FlicenseNot gradedqualityCmaintenanceEnables agents to query real-time OpenAPI documentation of backend services, providing tools to list services, endpoints, and schemas via MCP.
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to access a unified catalog of tools from various APIs (OpenAPI, GraphQL, MCP, Google Discovery) through the MCP protocol.MIT
- AlicenseNot gradedqualityBmaintenanceAn OpenAPI-driven MCP agent hub that dynamically generates MCP tools from Swagger/OpenAPI specs, enabling conversational interaction with backend APIs through natural language.14Apache 2.0
Related MCP Connectors
Operator-as-agent MCP hub. 6 tools. First $5 free, then $0.001/call.
Free public MCP for AI agents — 193 tools, 44 workflows. No API key.
Real-time Amazon, WIPO & PACER data for AI agents — 19 tools via the MCP protocol.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/dajarony/mcp-universal-gentes'
If you have feedback or need assistance with the MCP directory API, please join our Discord server