mcp-github-repo-creator

This MCP server provides a set of tools to analyze a local repository and automatically create a corresponding private GitHub repository, with topics and configuration. It exposes commands for analyzing repositories, generating a metadata file, creating a GitHub repository from that metadata, and performing a full end-to-end workflow. You can interact with these tools via natural language prompts in an MCP client (e.g., Copilot, Claude) or by invoking the underlying Python scripts directly. The available tools include: get_repo_analysis_instructions to request detailed analysis instructions, analyze_and_generate_metadata_file to produce github_repo_metadata.json, create_github_repo_from_metadata to create a repository from metadata, create_github_repository to create a repository from an existing metadata file, and full_repository_setup to run the complete analyze → create → connect workflow.

Once the server is running, you can connect via an MCP client using the provided server configuration, and issue the natural-language commands listed in the README. The workflow prioritizes private repositories, topic management, and automatic configuration so that the resulting GitHub repo mirrors the analyzed project locally, including topics and initial setup.

Prerequisites:

• Git, Python 3.x, and a POSIX shell (bash) installed on your system
• GitHub CLI (gh) for repository creation (optional but recommended)
• A GitHub account and a personal access token if you plan to push to GitHub programmatically

Steps:

1. Install dependencies and clone the project

   • Ensure Python is installed (python3 and pip)
   • Install Git if not present

   bash

2. Clone the repository and install the MCP server

   • git clone https://github.com/flickleafy/mcp-github-repo-creator.git
   • cd mcp-github-repo-creator
   • bash setup.sh

   This will create a Python virtual environment (venv), install the MCP SDK and dependencies, and prepare the project for use.

3. Run the MCP server

   • source venv/bin/activate
   • python server.py

4. Connect an MCP client

   • Use the MCP client configuration to connect to the server (see the README for the client JSON snippet). The server uses stdio transport and is compatible with Copilot, Claude Desktop, VS Code extensions, etc.

5. Optional: configure Copilot integration (VS Code example)

   • Ensure GitHub CLI is installed and authenticated (gh auth login)
   • Clone and configure as shown in the Copilot integration section of the README

Environment variable tips:

• If you run the server inside a virtual environment, ensure PATH includes the venv/bin directory when launching the MCP client (or set it in the client configuration, as shown in the README).
• For GitHub operations, you may need a GitHub token with repo scope configured in your environment or via the GitHub CLI authentication flow.

Common issues:

• If the server cannot access Git or GitHub, verify network access and that the local git repository is correctly initialized.
• Ensure the working directory (cwd) in the MCP client configuration points to the actual location of the mcp-github-repo-creator project.
• When using Copilot or Claude Desktop, ensure the PATH and venv are correctly exposed to the external tool via the provided env configuration.