Spec-Driven Development Guide

A comprehensive guide to systematic feature development using the three-phase spec process: Requirements โ Design โ Tasks.
Used by
- @kazini in their spec kit https://github.com/kazini/yask-spec-kit
- Amazon in internal presentations: "Itโs the best stuff I have found on [spec driven development]. I will be sharing links back to your repo with full attribution."
๐งญ Navigation Guide
New to spec-driven development? โ Start with Methodology Overview
Ready to create your first spec? โ Jump to Process Guide
Looking for examples? โ Browse Examples & Case Studies
Need templates? โ Get Ready-to-Use Templates
Working with AI? โ Learn Prompting Strategies
๐ Need detailed navigation? โ See Complete Navigation Index - Find content by role, problem, or learning style
๐ Complete Table of Contents
Learn the foundational concepts and philosophy behind spec-driven development
Step-by-step walkthrough of the three-phase workflow
Insights into decision-making frameworks and thought processes
Effective communication techniques for AI collaboration
Practical guidance for implementing features from specs
Curated references and learning materials
Real-world case studies and complete spec examples
Ready-to-use templates and checklists
Quick Start
New to spec-driven development? Start here:
- Understand the Methodology - Read the Overview to grasp core concepts
- See It in Action - Review a Simple Feature Spec example
- Try It Yourself - Use the Requirements Template for your first spec
- Get Better Results - Apply Prompting Strategies for AI collaboration
Navigation Tips
- ๐ Process sections provide step-by-step instructions
- ๐ง AI Reasoning sections explain the "why" behind decisions
- ๐ฌ Prompting sections help you communicate effectively with AI
- ๐ Examples show complete, real-world applications
- ๐ Templates give you ready-to-use starting points
๐ Cross-References & Related Content
By Workflow Phase
By Experience Level
Quick Problem Solving
๐ Kiro MCP Server
The Kiro MCP Server exposes Kiro's system prompts and instructions through the Model Context Protocol (MCP), allowing other AI assistants and tools to access Kiro's best practices.
Features
- Resources: Access all Kiro system documentation files via MCP resources
- Tools: Query and retrieve specific system instructions programmatically
- Prompts: Pre-configured prompts for common Kiro workflows
Quick Start
Install and configure the MCP server:
# Using uvx (recommended)
uvx kiro-mcp-server
# Or install with pip
pip install kiro-mcp-server
Add to your MCP client configuration (e.g., ~/.kiro/settings/mcp.json):
{
"mcpServers": {
"kiro-prompts": {
"command": "uvx",
"args": ["kiro-mcp-server"],
"disabled": false
}
}
}
For more details, see the MCP Server Documentation.
๐ฏ Claude Code Plugin
Kiro is available as an installable Claude Code plugin with 7 skills following the agentskills.io specification.
Quick Install
# In Claude Code
/plugin marketplace add https://github.com/jasonkneen/kiro
/plugin install kiro-spec-driven@kiro-marketplace
Available Skills
Once installed, Claude automatically uses these skills when relevant:
Installation Options
# User scope (available everywhere)
/plugin install kiro-spec-driven@kiro-marketplace
# Project scope (shared via git)
/plugin install kiro-spec-driven@kiro-marketplace --scope project
Validate Skills
./scripts/validate-skills.sh
For more details, see the Plugin Documentation and Skills Documentation.
This guide is designed to be both a learning resource and a reference manual. Jump to any section based on your current needs, or read through sequentially for comprehensive understanding.
๐ For detailed navigation by role, problem, or learning style, see the Complete Navigation Index