🚀 Configuration Guide: Spec Kit (GitHub)

Have you ever had a brilliant idea for a project but got stuck when trying to turn it into concrete tasks? Spec Kit, an open-source tool from GitHub, solves exactly this problem. It uses AI to convert natural language descriptions into technical specifications, implementation plans, and task backlogs — all ready to be executed.

In this guide, we’ll walk through the entire flow: from installation to automatic code generation.

🧩 Prerequisites

Before getting started, make sure you have:

• Python 3.10+ installed
• Node 18+ installed
• Git configured
• A GitHub account with Copilot access (recommended for the full workflow)
• Terminal with internet access

🛠️ Installation and Configuration

1. Install the UV Package Manager

uv is an ultra-fast Python package manager written in Rust. It replaces pip and virtualenv with vastly superior performance.

macOS (Homebrew):

brew install uv

macOS / Linux (script):

curl -LsSf https://astral.sh/uv/install.sh | sh

Windows (PowerShell):

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

💡 Tip: After installation, restart your terminal so the uv command is recognized.

2. Install Spec Kit

With uv ready, install the Spec Kit CLI directly from the official repository:

uv tool install specify-cli --from git+https://github.com/github/spec-kit.git

Verify the installation was successful:

specify version

3. Initialize the Project

Navigate to your project folder (or create a new one) and initialize the Spec Kit environment:

mkdir my-project && cd my-project
specify init .

This creates the .spec/ directory structure where specifications, plans, and tasks will be stored.

📝 Defining the Scope

4. Create the Specification

This is where the magic begins. Use the /specify command focused on the business goal (the “what”), without worrying about the technology (the “how”) yet. Be clear and descriptive:

/speckit.specify I want to build a web chat with generative AI. In this first version,
the system will be open (no login), with a modern, minimalist and fully
responsive design. The user will be able to send messages and receive
real-time responses with streaming.

📌 Best practices for specifications:

• Describe features from the user’s perspective
• Include important constraints (e.g., “no authentication”, “mobile-first”)
• Avoid mentioning technologies — that comes in the next step

5. Create the Technical Plan

Now it’s time to define the technological “recipe” using the /plan command. This is the moment to choose frameworks, languages, and architecture:

/speckit.plan The project will be a monorepo. Frontend: React + Vite, TypeScript, Shadcn UI
and TailwindCSS. Backend: Python with FastAPI, LangChain and LangGraph. LLM: OpenAI.
Real-time communication via Server-Sent Events (SSE).

Spec Kit will generate a detailed plan.md file with the proposed architecture, suggested directory structure, and dependencies.

🧐 Refinement and Execution

6. Validate the Plan with AI

Before jumping into code, ask Copilot (or your preferred LLM) to act as a Software Architect and audit the plan’s feasibility:

Audit the implementation plan and detail files.
Read with a critical eye to determine if:
1. There is a logical sequence of tasks
2. Dependencies between components are correct
3. There are technical risks or gaps in the specification

Reference the specific points in the implementation details
where crucial information is located.

⚠️ Important: This review step prevents rework. Take a few minutes to read the audit and adjust the plan before proceeding.

7. Generate Tasks (Backlog)

Transform the validated plan into a list of granular and actionable tasks:

/speckit.tasks

Each generated task will have:

• Clear description of what needs to be done
• Implicit acceptance criteria
• Execution order respecting dependencies

8. Automatic Implementation

Finally, ask Spec Kit to start generating code based on the validated plan:

speckit.implement .../.../plan.md

Spec Kit will create files, install dependencies, and structure the project as specified. Review the generated code and make adjustments as needed.

🎯 Flow Summary

💬 Final Thoughts

Spec Kit represents a significant shift in how we start software projects. Instead of spending hours manually structuring specifications and tasks, we can focus on what truly matters: the product vision.

Combine it with GitHub Copilot for an even more powerful workflow — from idea to working code in minutes.

Useful links:

• Spec Kit Repository on GitHub
• UV Documentation
• GitHub Copilot