# Spec Kitty

An open-source CLI that turns product intent into a repository-native, spec-driven workflow for AI coding agents, with isolated git worktrees and a kanban dashboard.

Spec Kitty is a command-line tool for spec-driven development with AI coding agents such as Claude Code, Codex, Cursor, Gemini, GitHub Copilot and Windsurf. It moves a project through a fixed sequence, from spec to plan to tasks to implementation to review to acceptance to merge, keeping specs, plans, work packages, acceptance criteria, review state and merge decisions inside the repository itself rather than in chat history.

Each agent gets an isolated git worktree so multiple agents or developers can implement work packages in parallel without branch conflicts, and a kanban dashboard tracks progress. It is meant for teams that are losing requirements or decisions across AI coding sessions, want reviewable work-package boundaries, and are moving from unstructured AI-assisted coding to a repeatable, git-tracked workflow; the authors describe it as unnecessary for one-off edits or teams that do not use git. It installs as a Python CLI and is licensed under MIT.

Website: https://spec-kitty.ai