husams/agile-mcp
If you are the rightful owner of agile-mcp and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
The Agile MCP Server is a specialized server designed to facilitate AI-driven agile software development lifecycle management.
Agile MCP Server
A Model Context Protocol (MCP) server that provides AI agents with structured tools for agile software development lifecycle management. This server enables AI assistants to autonomously manage backlogs, retrieve task-relevant information, and track generated artifacts in a structured, context-aware manner.
Overview
The Agile MCP Server addresses the need for a specialized tooling layer in AI-augmented development. Unlike traditional user-facing project management UIs, this MCP server provides granular, programmatic access to agile development processes that AI agents can interact with directly.
Key Features
- Backlog Management: Create, update, and organize user stories, epics, and tasks
- Artifact Tracking: Maintain context and relationships between generated development artifacts
- Story Workflows: Advanced workflow management for user story lifecycle
- MCP Protocol: Built on the Model Context Protocol standard for reliable agent-service communication
Architecture
This server implements the Model Context Protocol (MCP) specification, providing:
- Structured resource access for development artifacts
- Tool-based operations for backlog management
- Secure and predictable agent-service interactions
Version Information
Current Version: 2.0.0
ā ļø Breaking Changes in v2.0.0: All tool names have been updated to use underscore-based naming for better MCP protocol compliance. See the for complete details on updating from v1.x.
Getting Started
Prerequisites
- Python 3.8+
- uv package manager
Installation
-
Clone the repository:
git clone git@github.com:husams/agile-mcp.git cd agile-mcp
-
Install dependencies:
uv pip install -r requirements.txt
-
Configure MCP Client:
This MCP server uses stdio transport and must be configured in an MCP client (like Claude Desktop). It cannot run as a standalone process.
For Claude Desktop, add to your
claude_desktop_config.json
:{ "mcpServers": { "agile-mcp": { "command": "uv", "args": ["run", "run_server.py"], "cwd": "/path/to/agile-mcp" } } }
For other MCP clients:
- Transport: stdio
- Command:
uv run run_server.py
- Working Directory: Project root path
Usage
This MCP server provides tools and resources for:
- Managing agile backlogs and user stories
- Tracking development artifacts and their relationships
- Implementing structured agile workflows
Server Operation
Once configured in your MCP client, the server will automatically start when the client initializes. The server communicates via JSON-RPC over stdin/stdout as specified by the MCP protocol standard.
Project Structure
agile-mcp/
āāā src/agile_mcp/ # Main server implementation
ā āāā main.py # Server entry point
ā āāā api/ # API endpoints and handlers
ā āāā repositories/ # Data access layer
ā āāā services/ # Business logic
āāā docs/ # Project documentation
ā āāā prd.md # Product Requirements Document
ā āāā architecture/ # Architecture documentation
ā āāā prd/ # Detailed PRD sections
ā āāā stories/ # User story documentation
āāā tests/ # Test suite
ā āāā unit/ # Unit tests
ā āāā integration/ # Integration tests
ā āāā e2e/ # End-to-end tests
āāā requirements.txt # Python dependencies
Documentation
- - Complete guide for migrating from v1.x to v2.0
- - Version history and breaking changes
- - Complete list of MCP tools and their usage
Contributing
We welcome contributions! Please see our for details on our development process.
License
This project is licensed under the Apache 2.0 License. See the file for details.