task-master-memory-bank-mcp by divyad1508 - MCP Server

Task Master & Memory Bank MCP

The purpose of this document is to outline the setup and usage guidelines for the usage of task master & memory bank MCP. It will serve as the definitive source for developers whenever they encounter questions or issues regarding the utilization of the memory bank.

🚀 Features

Project Planning from a Specification: Intelligent Planning of the project to ensure the LLMs would work as you'd have if they were not present
Intelligent Context Execution: Smart project analysis and context-aware operations
Structured Memory Organization: Hierarchical template based memory bank system for different project aspects
MCP Integration: Full Model Context Protocol support for seamless AI assistant integration
Auto-Detection: Automatic project change detection and memory bank updates

📦 Setup Instructions (Approx. 5 - 10 minutes)

Prerequisites

Python 3.10 or higher
pip or uv package manager
Clone this repository in your local development environment

Step 1: Initialize & Activate a Virtual Environment

cd task-master-memory-bank-mcp
python3 -m venv .mcpvenv
source .mcpvenv/bin/activate

Step 2: Install Binaries for instantiating MCP Server

pip3 install -e .

Run following command to verify installation path:

which mcp-memory-bank

Sample Output should look like below:

/Users/divya-gangwar/Personal/task-master-memory-bank-mcp/.mcpvenv/bin/mcp-memory-bank

Step 3: Configure Cursor IDE to use MCP Server

PRE-REQUISITE: Please make sure to update your Cursor IDE to latest version

Navigate to any Cursor IDE Window, click on Cursor icon next to Apple icon on navigation top-bar. Select Settings → Cursor Settings -> Tools & Integrations.

On Tools & Integrations page, click on New MCP Server button. Clicking on it opens a mcp.json file, add the following JSON to this file.

{
  "mcpServers": {
    "memory-bank": {
      "command": "<PASTE the OUTPUT of which mcp-memory-bank from previous step HERE>",
      "args": [],
      "env": {},
      "description": "Intelligent memory management for AI assistants"
    }
  }
}

A sample JSON should look like below:

{
  "mcpServers": {
    "memory-bank": {
      "command": "/Users/divya-gangwar/Personal/task-master-memory-bank-mcp/.mcpvenv/bin/mcp-memory-bank",
      "args": [],
      "env": {},
      "description": "Intelligent memory management for AI assistants"
    }
  }
}

Once done, restart the Cursor IDE. After restart the Memory Bank MCP Server should be up & running in your local development environment.

For verifying, please run the following command in the terminal & you'll find a running MCP server process:

ps -ef | grep 'mcp'

Step 4: (CRITICAL for Memory Bank to work as expected) Memory Bank Workspace Configuration

Follow this basic-setup.md guide until Verification step.

Verify the configured environment variable is present across all terminal sessions. Sample output should look like below:

divya-gangwar@divya-gangwar ~ % echo "Projects: $PROJECT_WORKSPACES"
Projects: /Users/divya-gangwar/dev/memory_bank_poc,/Users/divya-gangwar/work

Please Note: PROJECT_WORKSPACES should contain Comma-separated list of all your project directories i.e. directories where you keep the projects in your local development environment.

Add the following in settings.json file of Cursor IDE. To navigate to this file, click on Cursor icon next to Apple icon on navigation top-bar. Select Settings → VS Code Settings -> Workbench -> Settings Editor -> Click on Edit in settings.json. It opens a JSON file, add/append the following JSON to that file.

"terminal.integrated.env.osx": {
    "ACTIVE_PROJECT_WORKSPACE": "${workspaceFolder}"
    "PROJECT_WORKSPACES": "<Add the output of echo $PROJECT_WORKSPACES from above>"
}

Sample Output should look like below:

"terminal.integrated.env.osx": {
    "ACTIVE_PROJECT_WORKSPACE": "${workspaceFolder}"
    "PROJECT_WORKSPACES": "/Users/divya-gangwar/dev/memory_bank_poc,/Users/divya-gangwar/work"
}

Once done, RESTART the Cursor IDE for all the changes to take effect.

At this point, you have completed the setup of Task Master & Memory Bank MCP. You can now go ahead & explore the power of the Specification Driven Development using Agentic AI & Memory Bank.

Usage Instructions

These instructions differ from the setup instructions, which were a one-time task. Adhering to these guidelines will enable you to maximize the utility of this MCP Server.

PLANNING PHASE

Define the Specification of the Task/Feature/Project you're going to build using this MCP Server

Write the project, feature, or task specification document in as much detail as possible to assist the agent in understanding the required tasks, the methods for execution, and the procedures for verifying the correctness of the generated work.

Initialization Prompt - Required only once in a repository

Before commencing work on a project/feature/task within a repository for the first time after the installation of the Memory Bank, please ensure that you utilize the following prompt to initialize the task list from the task specification document and to initialize the memory bank in the project.

Reference Prompt Template

Subsequent Usage Prompt - Use when the memory bank already exists in a repository

In the event that the project already includes the memory bank and you only need to generate the task list for your work, please utilize the following prompt.

Reference Prompt Template

EXECUTION PHASE

Following is a sample workflow of how you can make use of this framework once the installation & planning phase is complete for a feature:

Take a look at the technical_details.md file generated for each feature to ensure the Large Language Model (LLM) is going to work as you'd have liked to. If not, iterate upon the technical plan to arrive at a plan which aligns with your expectations. At this point, you should also make sure to respond to all the Review Questions in the technical_details.md file as it ensures the LLM is grounded to the context & is aligned to your plans.
You should also review the architecural decision logs & tasks file to make sure you're aligned with the breakdown of the work done by LLM.
Once above is done, you can ask the LLM to mark the feature as reviewed & ready for implementation by writing a sample prompt like below:
- Prompt: Mark <feature> as reviewed & ready for implementation
Once above is done, open the tasks.md file & start executing each phase/task 1 by 1.

For Upgrading Memory Bank MCP

After we publish an update, we will send a notification to everyone requesting the download and installation of the new binaries. This can be accomplished by executing the following commands in sequence after navigating to the directory where the memory bank source code is stored.

source .mcpvenv/bin/activate
git pull origin main
pip3 uninstall mcp-memory-bank
pip3 install -e .

Need Help

For any concerns, questions, or issues related to the usage of the memory bank, please reach out to us on email: divyad1508@gmail.com

🙏 Acknowledgments

Built with FastMCP
Inspired by the Model Context Protocol (MCP) specification
Thanks to the Cursor IDE team for excellent AI-assisted development tools

📊 Project Status

Version: 1.0.0
Status: Generally Available
Python: 3.10+
Maintainer: Swetank Shandilya, Divya Dyuti Gangwar