claude-code -management

This MCP server manager provides a command-line interface to install and configure Claude Code MCP servers for your user and project contexts. It exposes a set of predefined servers (for example, memory and github) that you can add, sync, list, and remove. The tool leverages the MCP server packages published under @anthropic-ai, which run specific Claude Code integrations or utilities as MCP servers. Use the add command to register servers to your local or project scope, and then use sync to apply the configured state. You can also operate on remote hosts or per-project configurations with the provided project- and remote- commands. The available servers in the configuration example include memory and github, each installed via npx and requiring appropriate environment variables when necessary (e.g., a GitHub token for the github server).

Prerequisites:

• Node.js 18+ (for Claude CLI and npx usage)
• Python 3.8+ with Ansible (pip install ansible)
• Claude Code CLI installed globally: npm install -g @anthropic-ai/claude-code

Installation steps:

1. Clone this repository: git clone https://github.com/your/repo.git cd repo
2. Install Python dependencies (Ansible): pip install ansible
3. Verify Claude CLI installation: claude --version
4. Prepare configuration: cp mcp-servers.json.example mcp-servers.json
5. Install/prepare MCP servers as needed via the manager commands (see Quick Start in README):
   • Add servers: make add SERVERS=memory,github
   • Sync all configured servers: make sync
   • List installed servers: make list

Notes:

• The actual MCP server packages (memory and github in the example) are installed via npx using the @anthropic-ai/mcp-server-* packages. Ensure you have network access to fetch these packages from npm.
• For servers requiring credentials, set them in a .env file or export environment variables before running the commands.

Tips and considerations:

• Environment variable substitution uses the ${VAR_NAME} syntax. Ensure you provide or export these values (for example, GITHUB_TOKEN) in your environment or a .env file.
• If you manage remote hosts, you can use the remote commands (sync-remote, add-remote) to configure MCP servers on another machine via SSH.
• Use the dry-run option to preview changes before applying: dry-run shows what would be added or removed without making changes.
• Review ansible/configuration guidance in docs/ADVANCED.md for advanced usage or custom playbooks.
• Always keep mcp-servers.json in sync with the actual server definitions you intend to manage; the example shows the structure for memory and github servers.