Claude Kiro: Spec-Driven Development for Claude Code
A unified CLI tool and methodology that brings spec-driven development workflow to Claude Code.
What This Is
Claude Kiro (ck) provides:
- Automated project setup for spec-driven development
- Smart hooks that provide spec context while coding
- Slash commands for structured workflows
- Output styles that enforce disciplined responses
Prompt → Requirements → Design → Tasks → Implementation
Repository Structure
├── VISION.md # Project vision and goals
├── synthesis/ # Implementation guides
│ ├── kiro-to-claude-mapping.md
│ └── phase1-implementation.md
├── research/ # Feature research from docs
│ ├── claude-code/ # Claude Code capabilities
│ └── kiro/ # Kiro methodology
└── resources/ # Scraped documentation
└── scraped/
├── docs.claude.com/ # 255 pages
└── kiro.dev/ # 89 pages
Quick Start
# Install Claude Kiro globally
uv tool install claude-kiro
# Initialize your project
cd your-project
ck init
# Verify setup
ck doctor
That's it! Your project is now configured for spec-driven development.
Installation
Install Claude Kiro (Global Tool)
# Install from PyPI
uv tool install claude-kiro
# Or install from source in editable mode
git clone https://github.com/angelsen/claude-kiro.git
cd claude-kiro
uv tool install . --editable
This installs the ck command globally, which provides:
ck init- Set up a project for spec-driven developmentck doctor- Verify your setup is workingck hook- Manage Claude Code hook integrationck --hook- Hook runner for Claude Code (hidden command)
Initialize Your Project
cd your-project
ck init
This creates:
.claude/output-styles/spec-driven.md- Enforces structured responses.claude/commands/spec/- Slash commands for specs.claude/settings.local.json- Hook configuration.claude/CLAUDE.md- Project context template
What the Hooks Do
The hooks provide intelligent spec context:
- When editing spec files: Shows which task you're implementing
- When editing new files: Suggests creating a spec first
- Smart caching: Shows messages only once per file per session (no spam!)
How It Works
- Initialize:
ck initsets up your project with all necessary files - Create specs:
/spec:create "feature"- Generate requirements, design, and tasks - Implement:
/spec:implement task- Execute tasks with full context - Track progress: TodoWrite tracks implementation automatically
- Stay aligned: Hooks provide context and maintain spec-driven discipline
CLI Commands Reference
Main Commands
ck init [--force]- Initialize a project with spec-driven setupck doctor- Check your Claude Kiro setup healthck hook list- Show available hooksck hook status- Display configured hooksck hook test <name>- Test a hook with sample datack hook config- Generate settings.json configuration
Claude Code Slash Commands (Created by ck init)
/spec:create <feature>- Create a new specification/spec:implement <task>- Implement a spec task/spec:review <spec>- Review an existing spec
Key Features
- EARS notation for testable requirements:
WHEN [condition] THE SYSTEM SHALL [behavior] - 3-phase workflow with approval gates between phases
- TodoWrite integration for native task tracking
- Smart hook context that tracks what you're working on
- Zero configuration after running
ck init
Documentation Sources
All research extracted from local scraped docs:
- Claude Code: 255 pages (docs.claude.com)
- Kiro: 89 pages (kiro.dev)
- Combined: 10 research docs, 2 synthesis guides
Implementation Status
- ✅ CLI tool (
ck) - Complete with all commands - ✅ Hook system - Smart context injection working
- ✅ Slash commands -
/spec:create,/spec:implement,/spec:review - ✅ Output styles - Spec-driven responses enforced
- ✅ Project setup automation -
ck initconfigures everything
Why This Exists
Problem: AI coding is fast but chaotic - implicit assumptions, undocumented requirements, hard to maintain.
Solution: Structured specs before code. Proven by Kiro, implemented in Claude Code.
Result: Production-ready development with AI assistance.
Built for developers who want structure without sacrificing speed.
Metadata
Release files for claude-kiro 0.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| claude_kiro-0.1.0.tar.gz | 22.8 kB | Without provenance | Details for claude_kiro-0.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| claude_kiro-0.1.0-py3-none-any.whl | Python 3 | none | any | Details for claude_kiro-0.1.0-py3-none-any.whl |
Total release size: 55.9 kB
Release files / claude_kiro-0.1.0.tar.gz
| Download URL | claude_kiro-0.1.0.tar.gz |
|---|---|
| Size | 22.8 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
84947dce22ca355a29d5b3dc75cdb7018c6d5ce27ba970ef428da1139f56d5f0
|
|
BLAKE2b-256 checksum How to use checksums |
ec413178f42ec0df293f500a90a7287e77277668e5f979697970a17ea1d4e087
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.8.22
|
Release files / claude_kiro-0.1.0-py3-none-any.whl
| Download URL | claude_kiro-0.1.0-py3-none-any.whl |
|---|---|
| Size | 33.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
9159ea7446e333d08f529eb18b202b9bd54dec8fbf30e0cacc5fe9f36ec82427
|
|
BLAKE2b-256 checksum How to use checksums |
f5cb2020e953bcc808dc0162fe9255fcf725642952fde5b5147d60b741169f21
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.8.22
|
Security for claude-kiro 0.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.