A curated list of awesome specification toolkits (speckits) for spec-driven development
Specification toolkits that drive AI to generate content. Define specs, let AI create code, docs, configs, tests, protocols, and more. Spec-Driven X (SD-X) for the AI era.
A speckit (specification toolkit) drives AI to generate content from specifications:
- 📝 Define specifications for your domain (API, protocols, configs, etc.)
- 🤖 AI Generates - Specs drive AI to create code, docs, configs, tests, designs, protocols
- ✅ Validate - Specs validate AI-generated content for quality and conformance
- 🎯 Consistency - Same spec ensures consistent outputs across the board
- 🚀 Automation - Specs automate what would be manual work
Spec-Driven X (SD-X) philosophy: Specs → AI → X (Anything)
Common examples include:
- SD-Development - Specs → AI → Code, tests, implementations
- SD-Design - Specs → AI → UI designs, architectures
- SD-Protocol - Specs → AI → Protocol definitions, standards
- SD-Documentation - Specs → AI → Docs, guides, tutorials
- SD-API - API specs → AI → Client SDKs, server implementations
- SD-Config - Config specs → AI → System configurations
- SD-Test - Test specs → AI → Test cases, scenarios
The "X" can be anything - these categories are not fixed. New SD-X applications emerge as the community discovers new ways to use spec-driven AI generation.
Speckits make specifications the single source of truth that drives AI.
How to build a speckit? Any way you want! Build from scratch, use a framework, or adapt existing tools - as long as it follows spec-driven principles. MetaSpec is one option if you want a framework, but you're completely free to implement your own solution in any language with any approach.
No speckits submitted yet - be the first to contribute!
# Install MetaSpec
pip install meta-spec
# Search for speckits
metaspec search "api validation"
# List all available
metaspec list
# Install a speckit
metaspec install <speckit-name># Install from PyPI
pip install <speckit-name>
# Use directly
<speckit-name> --helpHave a speckit to share? We'd love to have it!
1-Click Registration via Issue Template:
- Go to Create Registration Issue
- Fill in your speckit's GitHub repository URL
- Confirm you have permission to register
- Done! Our bot will:
- ✅ Automatically validate your speckit
- ✅ Extract metadata from
pyproject.toml - ✅ Create a PR automatically
- ✅ Notify you of the result
What the bot validates (automatically):
- Valid
pyproject.tomlwith name, version, description README.mdwith documentation- Working CLI commands in
[project.scripts] - Open source license
What you confirm: Permission to register/update this speckit
Already registered? Simply submit a new registration issue with the same repository URL. The system automatically:
- ✅ Detects it's an update (not a duplicate)
- ✅ Shows what changed (version, description, tags, etc.)
- ✅ Creates an update PR with clear changelog
See Update Guide for details.
If you prefer the traditional way:
# 1. Fork this repo
# 2. Edit speckits.json
# 3. Add your speckit entry
# 4. Create PRSee CONTRIBUTING.md for detailed guidelines.
We maintain quality to ensure good user experience:
- Clear, searchable descriptions
- Relevant tags
- Active maintenance
- Follows naming conventions
- GitHub Spec-Kit - Official Spec-Driven Development toolkit (46.6k+ stars)
- MetaSpec Framework - Meta-specification framework for creating speckits
- MetaSpec Documentation
- Creating a Speckit with MetaSpec
- Total Speckits: 0
- Contributors: 0
- Last Updated: 2025-11-09
This list is released under CC0 (Public Domain). Individual speckits have their own licenses.
Maintained by the ACNet-AI community and MetaSpec contributors.