baba786_phabricator-mcp-server

baba786_phabricator-mcp-server

3.4

Phabricator MCP Server provides an interface for large language models to interact with Phabricator APIs, focusing on task management, project, and user information. It is currently under development and requires a Phabricator API token for operation.

Phabricator MCP Server

A Model Context Protocol (MCP) server implementation for interacting with Phabricator API. This server allows LLMs to interact with Phabricator through a standardized interface.

Overview

This project provides an MCP server that exposes Phabricator functionality through:

  • Task management (viewing, creating, updating tasks)
  • Project information
  • User details

Getting Started

Prerequisites

  • Python 3.8+
  • Phabricator API token (from your Phabricator instance)
  • Access to a Phabricator instance

Installation

  1. Clone this repository:
git clone https://github.com/baba786/phabricator-mcp-server.git
cd phabricator-mcp-server
  1. Create and activate a virtual environment:
python -m venv venv
source venv/bin/activate  # On Unix/MacOS
# or
.\venv\Scripts\activate  # On Windows
  1. Install dependencies:
pip install -r requirements.txt
  1. Set up your environment:
# Copy the example env file
cp .env.example .env

# Edit .env and add your Phabricator token
# Replace 'your-token-here' with your actual Phabricator API token
echo "PHABRICATOR_TOKEN=your-token-here" > .env
  1. Run the server:
cd src
python server.py

Usage

Currently supported commands:

  • get-task: Retrieve details of a specific Phabricator task

Example usage through the client:

from src.mcp_minimal_client import Client

client = Client()
response = client.get_task(task_id="123")  # Replace with actual task ID
print(response)

Development Status

🚧 This project is currently under development. Stay tuned for updates!