Agile Planner MCP automates the creation of complete, structured agile backlogs directly within Windsurf, Cascade, or Cursor. It can generate epics, features, user stories, and even orphan stories from a simple description, with AI-generated annotations to guide implementation step-by-step. The server maintains a centralized backlog folder (.agile-planner-backlog) and supports intelligent feature–epic associations, progress tracking, and dependency management. When configured, you can trigger backlog generation via natural language prompts and then refine the output using the built-in validation and diagnostic logs to ensure the backlog conforms to the MCP schema.

With Agile Planner MCP, you’ll access a robust set of capabilities: automatic backlog structure creation (epics, features, user stories, and orphan stories), AI-optimized annotations that guide developers through tasks, and a clear directory organization for easy auditing. The server also provides diagnostic logs and structured error handling (e.g., consistent { success: false, error: { message: ... } } responses) to simplify debugging in your workspace environments like Windsurf, Cascade, or Cursor.

Prerequisites

• Node.js (14.x or newer) and npm installed on the host machine
• Optional: an OpenAI or GROQ API key if you plan to use AI capabilities

Install from npm (recommended for most users)

1. Install the MCP server package globally or in your project: npm install -g agile-planner-mcp-server

   or locally in your project:

   npm install agile-planner-mcp-server

2. Run the MCP server (example using npx): npx agile-planner-mcp-server

3. Configure in your workspace (example config): { "mcpServers": { "agile-planner": { "command": "npx", "args": ["agile-planner-mcp-server"], "env": { "MCP_EXECUTION": "true", "OPENAI_API_KEY": "sk-..." } } } }

Install from source (local development)

1. Clone the repository: git clone https://github.com/cyberlife-coder/agile-planner-mcp-server.git cd agile-planner-mcp-server

2. Install dependencies: npm install

3. Start the server in development mode: npm run start

4. Point your MCP config to the local server entry (path may vary): { "mcpServers": { "agile-planner": { "command": "node", "args": ["server/index.js"], "env": { "MCP_EXECUTION": "true", "OPENAI_API_KEY": "sk-..." } } } }

Notes and tips:

• Ensure OPENAI_API_KEY (or GROQ key) is provided if you intend to use AI-driven backlog generation. Keep keys secure and avoid committing them to VCS.
• The MCP output is organized under the .agile-planner-backlog folder for easy reference and auditing.
• If you encounter JSON schema validation errors, review the generated backlog structure against the MCP backlog format described in the documentation and check the diagnostic logs for clues.
• When using the npx option, you can leverage the latest published package without building locally, but for development or custom changes, running from source is recommended.