AI导哈

MCP服务器集合

用于Claude Code的个人MCP服务器集合，包含Jira、Confluence、GitHub、Ponymail和Slack的集成工具。

  • by cwlls
  • Python
  • 本地部署
  • 更新于
在 GitHub 查看 cwlls/MCPs

使用说明

单核细胞趋化蛋白

与Claude Code一起使用的MCP服务器的个人集合。

服务器

服务器 目录 工具
Jira jira/ search_issues, get_issue
汇流 confluence/ search_pages, get_page, list_spaces
GitHub github/ search_repos, get_repo, get_file, list_issues, get_issue, list_pull_requests, get_pull_request, search_code
Ponymail ponymail/ search_emails, get_email, get_list_stats, list_threads
Slack slack/ search_messages, list_channels, get_channel_history, get_thread, get_user

惯例

  • 运行时：Python 3.10+，使用管理 uv
  • 框架: FastMCP 来自 mcp.server.fastmcp
  • 运输标准： stdio
  • 身份验证/配置：仅环境变量
  • HTTP客户端: httpx （异步）

Claude代码集成

将服务器添加到您的设置中（全局 ~/.claude/settings.json 或项目 .claude/settings.json):

{
  "mcpServers": {
    "jira": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/jira/server.py"],
      "env": {
        "JIRA_BASE_URL": "https://jira.yourcompany.com",
        "JIRA_PAT": "your-token"
      }
    },
    "confluence": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/confluence/server.py"],
      "env": {
        "CONFLUENCE_BASE_URL": "https://confluence.yourcompany.com",
        "CONFLUENCE_PAT": "your-token"
      }
    },
    "github": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/github/server.py"],
      "env": {
        "GITHUB_PAT": "your-token"
      }
    },
    "ponymail": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/ponymail/server.py"],
      "env": {
        "PONYMAIL_BASE_URL": "https://lists.yourorg.org"
      }
    },
    "slack": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/slack/server.py"],
      "env": {
        "SLACK_TOKEN": "xoxp-your-token"
      }
    }
  }
}

设置

在使用任何服务器之前，请通过运行以下命令安装其依赖项 uv sync 在该服务器的目录中：

cd jira && uv sync
cd confluence && uv sync
cd github && uv sync
cd ponymail && uv sync
cd slack && uv sync

这将生成虚拟环境 uv run 当Claude创建服务器时使用。

Claude桌面集成（macOS）

编辑 ~/Library/Application Support/Claude/claude_desktop_config.json （如果不存在，则创建它）。格式相同 mcpServers 如上所述：

{
  "mcpServers": {
    "jira": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/jira/server.py"],
      "env": {
        "JIRA_BASE_URL": "https://jira.yourcompany.com",
        "JIRA_PAT": "your-token"
      }
    },
    "confluence": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/confluence/server.py"],
      "env": {
        "CONFLUENCE_BASE_URL": "https://confluence.yourcompany.com",
        "CONFLUENCE_PAT": "your-token"
      }
    },
    "github": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/github/server.py"],
      "env": {
        "GITHUB_PAT": "your-token"
      }
    },
    "ponymail": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/ponymail/server.py"],
      "env": {
        "PONYMAIL_BASE_URL": "https://lists.yourorg.org"
      }
    },
    "slack": {
      "command": "uv",
      "args": ["run", "python", "/path/to/MCPs/slack/server.py"],
      "env": {
        "SLACK_TOKEN": "xoxp-your-token"
      }
    }
  }
}

替换 /path/to/MCPs 带有此repo的绝对路径（例如。 /Users/yourname/MCPs).保存后，重新启动Claude Desktop以使更改生效。

添加新服务器

  1. 创建子目录： <name>/
  2. 加 server.py, pyproject.toml，以及 README.md
  3. 使用 @mcp.tool() 暴露工具；通过env-vars接受所有配置
  4. 跑 uv lock 生成锁文件
  5. 更新此自述文件 CLAUDE.md

相关 MCP