Spec Kitty: Spec-Driven Development for AI Coding Agents and Software Factories

Introduction

Spec Kitty is an open-source command-line interface (CLI) designed to bring structure and governance to AI-assisted software development. It facilitates spec-driven development, transforming product intent into a repeatable, repo-native AI coding workflow. By integrating with various AI coding agents like Claude Code, Cursor, Gemini, and GitHub Copilot, Spec Kitty helps teams build robust software factories with clear processes.

Why Use It and Key Features

Spec Kitty addresses common challenges in AI coding sessions, such as losing requirements, decisions, or acceptance criteria. It's ideal for teams moving from ad hoc "vibe coding" to a more structured, repeatable development workflow.

Key features include:

• Structured Workflow: Guides users through spec -> plan -> tasks -> next -> review -> accept -> merge lifecycle.
• Repository-Native Artifacts: Stores specs, plans, work packages, and review states directly in your Git repository under kitty-specs/.
• Parallel Development: Utilizes isolated Git worktrees under .worktrees/ to allow multiple agents or developers to work in parallel without branch chaos.
• Visible Progress: Offers a local Kanban dashboard (spec-kitty dashboard) for tracking mission progress.
• Governance and Audit Trail: Provides review, accept, and merge gates, ensuring human oversight and creating an audit trail.
• Multi-Agent Support: Integrates with a wide range of AI coding agents, including Claude Code, Codex, Cursor, Gemini, GitHub Copilot, and more.
• Retrospective Learning: Generates retrospectives for completed missions to foster continuous improvement.

Installation

Spec Kitty is recommended to be installed using pipx to ensure it runs in its own virtual environment, avoiding potential dependency conflicts.

To install the CLI:

pipx install spec-kitty-cli

Alternatively, you can use uv or pip within an activated virtual environment:

uv tool install spec-kitty-cli
# or, inside an activated virtual environment
python -m pip install spec-kitty-cli

After installation, initialize a new project:

spec-kitty init my-project --ai claude
cd my-project
spec-kitty verify-setup

Replace claude with your preferred agent key, such as codex, cursor, gemini, or copilot.

Examples

Here's a quick example of the core workflow using Spec Kitty:

1. Charter and Specify: Define the mission and its requirements.

   /spec-kitty.charter
   /spec-kitty.specify Build a small task list app.
   

2. Plan and Tasks: Create a plan and break it down into tasks.

   /spec-kitty.plan
   /spec-kitty.tasks
   

3. Execute Next Action: Let Spec Kitty determine the next step for the agent.

   spec-kitty next --agent claude --mission <mission-slug>
   

4. Review and Merge: Review the agent's work, accept it, and merge.

   /spec-kitty.review
   /spec-kitty.accept
   /spec-kitty.merge --push
   

For a full walkthrough, refer to the official documentation.

Links

• GitHub Repository: https://github.com/spec-kitty/spec-kitty
• PyPI: https://pypi.org/project/spec-kitty-cli/
• Documentation: https://github.com/spec-kitty/spec-kitty/blob/main/docs/guides/tutorials/getting-started.md