MCP.so
ログイン

BaseMcpServer

@dawsonlp

BaseMcpServer について

A minimal containerized base for MCP servers using the MCP Python SDK

基本情報

カテゴリ

その他

ランタイム

python

トランスポート

stdio

公開者

dawsonlp

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "BaseMcpServer": {
      "command": "uv",
      "args": [
        "tool",
        "install",
        "git+https://github.com/dawsonlp/BaseMcpServer.git#subdirectory=utils/mcp_manager"
      ]
    }
  }
}

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is BaseMcpServer?

A monorepo of Model Context Protocol (MCP) servers plus the mcp-manager CLI, which installs and configures them. It includes servers for Jira/Confluence integration, current-context tools, and a server generator, all sharing a common infrastructure.

How to use BaseMcpServer?

Install uv and mcp-manager via a one‑line curl command, then use mcp-manager install local <server> --source <path> to install individual servers. After editing per‑server config files, run mcp-manager config sync to wire all servers into VS Code/Cline and Claude Desktop over stdio.

Key features of BaseMcpServer

  • mcp-manager CLI for installing, managing, and configuring MCP servers
  • Each server runs in its own isolated uv‑managed virtual environment
  • Single config sync command updates both Cline and Claude Desktop
  • Multi‑instance Jira + Confluence server (jira-helper)
  • Current‑context tools: date/time, market, news, dev‑tool versions (worldcontext)
  • Server generator that scaffolds new MCP servers from a code snippet (mcpservercreator)
  • Starter template for building custom servers with shared infrastructure

Use cases of BaseMcpServer

  • Deploy and manage multiple MCP servers across editors from a single CLI
  • Automate Jira and Confluence workflows via an LLM‑connected MCP server
  • Provide an AI agent with current context such as date, market data, and news
  • Generate a new MCP server on‑the‑fly from a code snippet for prototyping
  • Fork the template to build a new server with robust tool registration and logging

FAQ from BaseMcpServer

What does mcp-manager do?

It installs MCP servers from local sources, manages their isolated uv environments under ~/.config/mcp-manager/servers/<name>/.venv, and writes mcpServers entries to both Cline and Claude Desktop configuration files via mcp-manager config sync.

How are servers installed and configured?

First install mcp-manager with uv tool install, then run mcp-manager install local <server> --source <path> for each server. Edit the per‑server config file (e.g., ~/.config/mcp-manager/servers/jira-helper/config.yaml) to add credentials, then run mcp-manager config sync to enable the servers in your editors.

Does config sync overwrite my hand‑added editor entries?

No. Only entries whose server name matches a server in the mcp‑manager registry are overwritten. Hand‑added entries in your editor’s settings file are preserved across config sync.

What runtime does BaseMcpServer require?

Python and uv. Each server runs as a Python package inside its own uv‑managed virtual environment.

How do I connect a client after installing servers?

Run mcp-manager config sync to write stdio‑based mcpServers entries to both Cline and Claude Desktop. Restart your editor for the new entries to take effect.

コメント

「その他」の他のコンテンツ