Skillget

haive-mcp

mcp
Guvenlik Denetimi
Uyari
Health Uyari
  • No license — Repository has no license file
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 9 GitHub stars
Code Gecti
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Gecti
  • Permissions — No dangerous permissions requested
Purpose
This tool acts as a dynamic integration hub for AI agents, allowing them to automatically discover, install, and connect to over 1,960 external servers via the Model Context Protocol (MCP). It bridges the gap between AI frameworks like LangChain and external tools, offering containerized isolation and CLI management.

Security Assessment
The overall risk is rated as Medium. While the automated code scan (covering 12 files) found no dangerous patterns or hardcoded secrets, the tool's core functionality inherently involves executing external shell commands (like running `npx` or Docker containers) and making network requests to discover and fetch third-party packages. The application mitigates some of this exposure through a Human-in-the-Loop (HITL) approval system for installations and offers Docker isolation. No excessive system permissions are required, but because it dynamically pulls and runs external code, standard sandboxing precautions are highly recommended.

Quality Assessment
The project is very new and currently has low community visibility with only 9 GitHub stars. However, it shows strong active maintenance, with repository updates pushed as recently as today. There is a minor discrepancy in its health check: the repository lacks an actual license file in its directory, despite claiming to be MIT licensed via its README badge. The development structure appears professional, featuring automated CI pipelines and hosted documentation.

Verdict
Use with caution — the underlying code is safe, but because the tool dynamically fetches and executes third-party integrations, you should only run it in isolated environments.
SUMMARY

Dynamic MCP integration for AI agents — search 1,960+ servers, install with HITL approval

README.md

haive-mcp

PyPI version
Python Versions
License: MIT
CI
Docs
PyPI Downloads

Dynamic MCP integration for AI agents — search 1,960+ servers, install with HITL approval, connect via LangChain or Docker.

haive-mcp enables runtime discovery and integration of tools from the Model Context Protocol (MCP) ecosystem. Agents can find and use tools from 1,960+ MCP servers without predefined configuration.

Installation

pip install haive-mcp

Features

  • 🔍 Server Discovery — search across 1,960+ MCP servers
  • 🚀 Multi-Transport — STDIO, SSE, Streamable HTTP, Docker
  • 🛡️ HITL Approval — Human-in-the-loop for server installation
  • 🤖 Intelligent Agent — IntelligentMCPAgent auto-discovers tools for tasks
  • 🐳 Docker Isolation — run servers in containers for security
  • 📦 LangChain Bridge — MCP tools work as standard LangChain tools
  • ⚡ CLI — haive-mcp command for discovery and management

Quick Start

Static Configuration

from haive.mcp.config import MCPConfig, MCPServerConfig, MCPTransport

config = MCPConfig(
    enabled=True,
    servers={
        "filesystem": MCPServerConfig(
            name="filesystem",
            transport=MCPTransport.STDIO,
            command="npx",
            args=["-y", "@modelcontextprotocol/server-filesystem"],
        )
    }
)

Dynamic Discovery

from haive.mcp.agents import IntelligentMCPAgent
from haive.core.engine.aug_llm import AugLLMConfig

agent = IntelligentMCPAgent(
    engine=AugLLMConfig(),
    auto_discover=True,
    require_approval=True,
)

result = await agent.arun("Search GitHub for Python repos about quantum computing")
# Auto-discovers and installs github-mcp-server, then uses it

CLI

# Discover servers for a topic
haive-mcp discover "database"

# List available transports
haive-mcp transports

# Install a server
haive-mcp install postgres

Transport Types

Transport Use Case
stdio CLI servers via npx/uvx (most common)
sse HTTP streaming servers
streamable_http Continuous data transfer
docker Isolated container execution

Documentation

📖 Full documentation: https://pr1m8.github.io/haive-mcp/

Related Packages

Package Description
haive-core Foundation: engines, graphs
haive-agents Production agents
haive-tools Static tool implementations

License

MIT © pr1m8

Yorumlar (0)

Sonuc bulunamadi