spec-coding

Spec Coding MCP Server enables an IDE-driven workflow that connects external tooling with an AI-assisted development assistant. It centers around the concept of a Spec folder for each functional module, containing requirements.md, design.md, and tasks.md to drive the development process from specification through task execution. You’ll interact with the MCP server through your editor or IDE integration, which reads the server’s guidance and translates it into concrete development steps. The server exposes capability to generate and manage EARS-formatted requirements, architectural design notes, and a runnable task list, enabling a traceable, repeatable spec-to-code pipeline that supports AI-assisted coding workflows.

To use it, start the MCP server in your chosen environment (in this case via Docker). Once running, configure your IDE (as described in the project docs) to point at the MCP server. The workflow typically proceeds through: (1) Function Definition in your editor to initiate the spec-coding process, (2) Requirements Collection where the assistant generates an EARS-formatted requirements document, (3) Design Document creation outlining architecture and approach, (4) Task Planning to decompose the solution into executable steps and create the Spec folder structure, and (5) Task Execution where the assistant or your team works through the generated tasks until development is complete. The tools provided by the server help maintain rigor and traceability across the spec-driven development lifecycle.

Prerequisites:

• Docker installed and running on your machine
• Basic familiarity with using MCP-enabled IDE integrations (e.g., VS Code with the MCP extension)

Installation steps:

1. Pull and run the MCP server via Docker:

   docker run -i feiyun0112/spec-coding-mcp:latest

2. If you prefer a local environment without Docker, obtain the application package from the project repository (for example via NuGet if you switch to a .NET runtime) and run the server according to the project’s local setup instructions (noting that this project’s README indicates a Docker-based deployment as the recommended route).

3. In your IDE, configure the MCP connection to point to the running server. The configuration steps are typically exposed in the IDE’s MCP settings (e.g., a .vscode/mcp.json file) as shown in the project’s guidance.

Notes and tips:

• This MCP server is described as a Spec-Coding MCP Server, intended to support Spec-Driven Development with a focus on Requirements (EARS syntax), Design documents, and Task planning/execution.
• The project references .NET 10 as a running requirement in its documentation, but the Docker-based deployment shown here is a portable alternative. If you run locally, ensure compatibility with your environment’s runtime and dependencies.
• If you encounter issues starting the MCP, verify Docker is functioning, check network/port accessibility for the IDE integration, and confirm that the container image feiyun0112/spec-coding-mcp:latest is accessible (pull it if necessary).
• Environment variables are often optional or placeholder-based for MCP setups; adjust as needed for your environment (e.g., timeouts, logging levels) depending on your IDE integration and container configuration.