A specification-first development framework with live kanban dashboard and AI agent orchestration, built on GitHub's Spec Kit.
Key Features:
- 📊 Live Kanban Dashboard – Real-time visibility into work across planned → doing → for review → done lanes
- 🔄 Multi-Agent Orchestration – Coordinate multiple AI coding agents (Claude Code, GitHub Copilot, Gemini CLI, Cursor, Windsurf, and more)
- 🎯 Spec-Driven Development – Flip the traditional model: specifications become executable, directly generating working implementations
- 📦 Artifact Management – Track specifications, plans, tasks, and deliverables in one integrated workspace
- 🔧 Agent-Aware Prompts – Scaffolding commands tuned to each AI agent's capabilities
- ⚡ Zero Configuration – Automated dashboard starts with
spec-kitty init - 🌳 Worktree Strategy – Isolated sandboxes for parallel feature development
What it does: Spec Kitty changes how teams build software by emphasizing specification-first rigor. Instead of treating specs as throwaway documents, they become the source of truth that drives implementation. The built-in dashboard gives you real-time insights into your AI-assisted development workflows, showing exactly which agents are working on what and how tasks move through your kanban board.
Workflow:
- Constitution – Establish project principles and governance
- Specify – Define requirements and user stories
- Clarify – Structured discovery interviews to reduce ambiguity
- Plan – Create technical implementation plans with your tech stack
- Tasks – Break down into actionable work packages
- Implement – Execute with AI agent assistance
- Review – Validate and move to done
- Accept & Merge – Finalize and integrate
Supported AI Agents:
- Claude Code
- GitHub Copilot
- Gemini CLI
- Cursor
- Windsurf
- Qwen Code
- opencode
- Amazon Q Developer CLI
- Kilo Code
- Auggie CLI
- Roo Code
- Codex CLI
Installation:
# From PyPI (Recommended)
pip install spec-kitty-cli
# Or with uv
uv tool install spec-kitty-cli
# Initialize a new project
spec-kitty init my-project --ai claudeRepository: github.com/Priivacy-ai/spec-kit
- Read the Spec Kitty README
- Run
spec-kitty --helpfor CLI reference - Check playbook examples:
- Multi-agent feature development
- Parallel implementation tracking
- Dashboard-driven development
- Claude + Cursor collaboration
Spec Kitty is released under the MIT License.
- Report Issues: Spec Kitty Issues
- Website: priivacy.ai
- Discussions: Check GitHub Discussions in each repository
We welcome contributions! Please open an issue or pull request in the relevant repository. Be sure to follow the project's code of conduct and contribution guidelines.
Built with ❤️ by the Priivacy AI team