One Open Source Project a Day (Part 55): Spec Kit - GitHub's Open-Source Spec-Driven Development Toolkit

Introduction

"Build high-quality software faster. Focus on product scenarios and predictable outcomes instead of vibe coding every piece from scratch."

This is Part 55 of the "One Open Source Project a Day" series. Today's project is Spec Kit (GitHub).

Tired of vibe coding from scratch every time? Spec Kit is GitHub's open-source Spec-Driven Development toolkit: it makes specifications executable, directly generating working implementations instead of just guiding them. It provides a structured workflow constitution → specify → plan → tasks → implement, with Specify CLI and /speckit. slash commands*, supporting Claude Code, Cursor, GitHub Copilot, Gemini CLI and 20+ AI coding assistants.

What You'll Learn

Spec-Driven Development philosophy
Spec Kit's five-step workflow
Specify CLI and slash commands
Supported AI coding assistants
Quick start and typical usage

Prerequisites

Basic understanding of AI-assisted programming
Experience with Claude Code, Cursor, or similar tools (optional)

Project Background

Project Overview

Spec Kit is GitHub's open-source Spec-Driven Development toolkit. In traditional development, specs are scaffolding and code is king; Spec-Driven Development makes specs executable, directly generating implementations. Spec Kit provides a structured flow: establish project principles (constitution), define requirements (specify), create technical plan (plan), break down tasks (tasks), then execute implementation (implement).

Team

Team: GitHub
Philosophy: Intent-driven, spec-first, multi-step refinement, leveraging advanced AI capabilities

Project Data

📄 License: MIT
🌐 Website: github.github.com/spec-kit
📚 Documentation: GitHub README
💬 Community: GitHub Issues

Core Features

Main Workflow

Step	Command	Description
1. Principles	`/speckit.constitution`	Create project governing principles and dev guidelines
2. Requirements	`/speckit.specify`	Describe what to build (what/why), not tech stack
3. Technical Plan	`/speckit.plan`	Specify tech stack and architecture, generate plan
4. Task Breakdown	`/speckit.tasks`	Generate actionable task list from plan
5. Implementation	`/speckit.implement`	Execute tasks in order, generate code

Optional commands: /speckit.clarify (clarify requirements), /speckit.analyze (cross-artifact consistency), /speckit.checklist (quality checklist)

Use Cases

Greenfield (0-to-1): Generate spec, plan, tasks, and implementation from scratch
Brownfield extension: Iteratively add features, modernize existing codebases
Creative exploration: Explore multiple tech stacks and architectures in parallel

Quick Start

Install Specify CLI:

# Recommended: persistent install
uv tool install specify-cli --from git+https://github.com/github/spec-kit.git
 
# Or one-time run
uvx --from git+https://github.com/github/spec-kit.git specify init my-project

Initialize project:

# New project
specify init my-project --ai claude
 
# Initialize in current directory
specify init . --ai claude
# or
specify init --here --ai cursor-agent
 
# Check environment
specify check

Use slash commands in AI assistant:

/speckit.constitution Create principles focused on code quality, testing standards, UX consistency
 
/speckit.specify Build an application that helps organize photos in albums by date, with drag-and-drop reordering
 
/speckit.plan Use Vite with vanilla HTML/CSS/JS, local SQLite for metadata
 
/speckit.tasks
 
/speckit.implement

Supported AI Assistants

Officially supported: Claude Code, Cursor, GitHub Copilot, Gemini CLI, Codex, Qwen Code, Windsurf, Kiro CLI, Mistral Vibe, Kimi Code, and more

Generic mode: --ai generic --ai-commands-dir <path> for custom assistants

Project Advantages

Comparison	Spec Kit	Pure Vibe Coding	Traditional PRD
Spec role	Executable, generates implementation	None	Reference only
Workflow	Structured five-step	None	Document-centric
Tech stack	Specifiable, explorable	Ad-hoc	Separate design
Task dependencies	Auto-ordered	None	Manual
AI integration	20+ assistants	Varies	None

Deep Dive

Development Phases

Phase	Focus	Activities
0-to-1 (Greenfield)	Generate from scratch	High-level requirements → spec → plan → implementation
Creative exploration	Parallel implementations	Multiple tech stacks, architectures, UX patterns
Iterative enhancement (Brownfield)	Brownfield modernization	Add features iteratively, modernize legacy systems

Directory Structure (after init)

.specify/
├── memory/
│   └── constitution.md      # Project principles
├── specs/
│   └── 001-feature-name/
│       ├── spec.md           # Feature spec
│       ├── plan.md           # Implementation plan
│       ├── tasks.md          # Task list
│       ├── data-model.md     # Data model
│       └── ...
├── scripts/                  # Scripts
└── templates/                # Templates

Experimental Goals

Technology independence: Validate that Spec-Driven Development is not tied to specific tech
Enterprise constraints: Support cloud providers, tech stacks, compliance
User-centric development: Adapt to different user cohorts and dev approaches
Creative & iterative: Support parallel exploration and iterative enhancement

Project Links and Resources

Official Resources

🌟 GitHub: https://github.com/github/spec-kit
🌐 Website: github.github.com/spec-kit
📚 Documentation: README, spec-driven.md
💬 Issues: GitHub Issues

Target Audience

Developers wanting less vibe coding and more predictable outcomes
Teams needing structured AI development workflows
Teams with both greenfield and brownfield projects
Users of Claude Code, Cursor, Copilot, and similar AI coding assistants

Visit my homepage for more useful knowledge and interesting products