git-context-mcp runs entirely locally and exposes read-only Git-informed context to AI agents over MCP. It analyzes the repository by inspecting Git metadata and repository files to provide structured insights such as the repository state, entry points, recent activity, maintenance risk, and overall code structure. Tools are designed to be consumed by MCP-aware clients via STDIO, allowing agents to query the server for concrete context before making editing or review decisions. The included tools focus on high-signal information rather than raw code, helping AI assistants understand where to start, what changed recently, and which files carry higher risk.

To use the server, start the MCP process using the recommended Python path and run it in a local environment where the target Git repository exists. Once running, you can invoke the exposed tools (project_status, code_map, recent_activity, work_summary, risk_scan) to retrieve structured insights. Each tool is read-only and tailored to answer targeted questions like current repository state, entry points, recent churn, and debt indicators. This makes it easier for an AI agent to reason about changes, plan edits, and avoid risky areas without needing to access or distribute the codebase beyond the analysis results.

Prerequisites

• Python 3.10 or newer
• Git installed and available on PATH
• A local clone of the repository to analyze

Installation steps

1. Clone the repository and navigate into it git clone https://github.com/TamiShaks-2/git-context-mcp.git cd git-context-mcp

2. Create and activate a virtual environment python -m venv .venv

   macOS / Linux

   source .venv/bin/activate

   Windows

   .venv\Scripts\activate

3. Install the MCP package in editable mode pip install -e .

4. (Optional) Verify installation by listing available scripts or running a minimal test if provided in the repo pytest

5. Run the MCP server as described in the README to begin serving context via STDIO

Tips and common considerations:

• The server is read-only and does not modify the repository; it solely analyzes Git metadata and files.
• Ensure the Python environment has access to the target Git repository location.
• When using MCP Inspector or other clients, the server will respond to the tools in the following order: project_status, code_map, recent_activity, work_summary, risk_scan.
• If you encounter environment issues, ensure Python 3.10+ is active in your terminal and that you’ve activated the virtual environment before running the server.
• If the repository is large, consider working within a shallow clone or limiting history depth to speed up analysis.
• No external services or API keys are required; all data is read from the local Git repository.

Configuration options (beyond defaults):

• You can extend or customize the server by editing the src/tools modules (e.g., project_status.py, code_map.py) to adjust the scope of analysis or the formatting of returned context.