spec-workflow

Spec Workflow MCP is a server that manages a structured spec-driven development workflow with a real-time dashboard and optional VSCode integration. It orchestrates a sequential flow from requirements to design to tasks, provides a live dashboard to monitor specs, progress, and implementations, and supports an integrated VSCode extension for editing and monitoring right inside your editor. Typical usage involves starting the MCP server and connecting a dashboard interface to manage and track specs across multiple projects. You can issue commands like creating a complete spec workflow, listing existing specs, or executing individual tasks within a spec, all while the dashboard updates in real time as work progresses.

To use the tools, add the spec-workflow MCP server to your client configuration, then choose your interface: the Web Dashboard for a centralized view of all projects (default port is 5000), or the VSCode Extension for a tightly integrated experience in your editor. The dashboard exposes views for specs, tasks, progress bars, and implementation logs, while the VSCode extension provides an additional sidebar dashboard and quick-access commands. Examples include creating a spec for a feature, listing specs to monitor their status, and executing a specific task like "Execute task 1.2 in spec user-auth". See the docs for more prompting examples and integration guidance.

Prerequisites:

• Node.js and npm (or pnpm/yarn as preferred)
• A project path to pass to the MCP server

Installation steps:

1. Install Node.js from the official website if you don't have it already.

2. Ensure you have npx available (comes with npm):

   • Verify: node -v npm -v npx -v

3. Start the MCP server via npx using the recommended package scope:

   • Quick start (as shown in the docs):

     npx -y @pimzino/spec-workflow-mcp@latest /path/to/your/project

   • If you want the web dashboard, start the dashboard separately (default port 5000):

     npx -y @pimzino/spec-workflow-mcp@latest --dashboard

4. (Optional) If you prefer Docker deployment, follow the Docker deployment guide in the README (build and run the container, then access the dashboard at http://localhost:5000).

5. Add to MCP client configuration where your tools read MCP settings, using the provided mcp_config structure.

Tips and notes:

• The -y flag bypasses npm prompts for smoother installation when using npx.
• When using the dashboard with multiple projects, note that only one dashboard instance is needed as all projects connect to the same dashboard.
• If you use the CLI or integration clients (Claude, Cline/Claude Dev, Continue IDE, Cursor IDE, OpenCode, Windsurf, Codex, etc.), configure the mcpServers entry with command: "npx" and args: ["-y", "@pimzino/spec-workflow-mcp@latest", "/path/to/your/project"].
• For Windows or environments with path handling issues, you may need to adapt the command invocation accordingly (e.g., using cmd.exe on Windows).
• Docker deployment provides isolation and reproducible environments; the dashboard runs on port 5000 by default.
• The MCP supports a structured workflow, real-time updates, and an approval/configuration model through the dashboard and VSCode extension.