spec-coding-mcp
Transform feature ideas into production-ready code through systematic Spec-Driven Development 通过系统化的**规格驱动开发**,将功能想法转化为可投入生产的代码
Install / Use
claude mcp add feiyun0112 -- npx -y github:feiyun0112/spec-coding-mcpIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
Development & EngineeringSupported Platforms
Tags
Skill content
View source on GitHubSpec Coding MCP Server
Transform feature ideas into production-ready code through systematic Spec-Driven Development
What is Spec-Driven Development?
Spec-Driven Development is, simply put, a software development process where clear, detailed specifications serve as the core driver throughout the entire development workflow. These specifications cover various aspects such as functional requirements, performance metrics, interface definitions, and data formats of the software system, acting like blueprints in construction to clearly outline what the final software should look like for developers.
The core of spec-driven development is to transform vague requirements into an executable, traceable, and verifiable development process through standardized specification documents. Essentially, it integrates the rigor of traditional software engineering into AI programming, turning development from "vibe coding" (relying on intuition) into a controllable engineering process.
The key to spec-driven development is the "Spec folder" — each functional module corresponds to a Spec folder containing 3 core files that "materialize" requirements, designs, and tasks:
-
1.requirements.md (Requirements Document) — Uses EARS syntax (Easy Approach to Requirements Syntax) to write user stories and acceptance criteria, avoiding vague expressions.
-
2.design.md (Technical Solution) — Includes details such as architectural design, process logic, technology selection, and potential risks.
-
3.tasks.md (Task List) — Breaks down the technical solution into executable specific tasks (todolist).
Implementing Spec-Driven Development with spec-coding-mcp
The MCP in
spec-coding-mcpstands for Model Context Protocol, which is an important bridge connecting external tools with the AI IDE. As long as your AI IDE can use MCP, you can usespec-coding-mcp.
Here, we'll take GitHub Copilot as an example to introduce how to use spec-coding-mcp.
1. Configure Vs Code
To make GitHub Copilot work with Spec-Coding-MCP, you first need to set up MCP server information in the relevant configuration files.
Visit NuGet.org, search for SpecCodingMcpServer, and you'll see the "MCP Server" tab:

Add a .vscode/mcp.json file to your Vs Code workspace, copy the json from the "MCP Server" tab into it, and save. Then click start above:

SpecCodingMcpServer requires .NET 10 to run. Please ensure you have .NET 10 installed locally (https://dotnet.microsoft.com/en-us/download/dotnet/10.0)

2. Start the Spec-Driven Development Process
- 1.Function Definition: Issue the instruction "Start spec coding" to GitHub Copilot, then specify the specific function, such as "Create a Vue to-do app", add relevant details, and proceed to the next step after confirmation.



- 2.Requirements Collection: Copilot will start the requirements collection phase, generate a requirements document in EARS format, and proceed to the design phase after confirming the requirements are correct.

- 3.Design Document: Based on the confirmed requirements, Copilot will generate a comprehensive technical architecture design document. Proceed to task planning after confirming the design is acceptable.

- 4.Task Planning: Copilot will decompose the design and requirements into a list of executable tasks, ultimately generating a complete Spec folder structure containing requirements.md, design.md, and tasks.md.

- 5.Task Execution: Enter the specific execution phase after confirming the task plan. Copilot will execute tasks one by one until development is complete.




Related Skills
Agent-Reach
84.2kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
headroom
73.4kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
ruflo
73.0k🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
career-ops
72.3kOpen-source AI job search: scan job portals, evaluate listings into a structured A-H report with a global 1-5 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)
