spec-kitty

Spec Kitty is an open-source CLI for turning product intent into a repo-native AI coding workflow:

spec -> plan -> tasks -> next -> review -> accept -> merge

Use it to build a governed software factory around Claude Code, Codex, Cursor, Gemini, GitHub Copilot, Windsurf, OpenCode, and other AI coding agents. Spec Kitty keeps specs, plans, work packages, acceptance criteria, review state, and merge decisions in your repository, then gives agents isolated git worktrees so implementation can happen in parallel without branch chaos.

Latest release candidate: 4.0.0rc8

main carries the 4.x release-candidate line. To try the latest candidate, install its exact version:

uv tool install 'spec-kitty-cli==4.0.0rc8'

Then run spec-kitty upgrade in each project and commit the result. What changed and what to do is in the 4.0.0rc8 release notes.

Release candidates are for qualification; stable launch acceptance is still pending. A normal install without a version still selects the latest stable release.

Bright Software Factory, Not a Black Box

Spec Kitty is for teams building software factories: repeatable inputs, clear work-package boundaries, isolated execution, visible progress, and review gates. It can support dark software factories and autonomous coding experiments, but it is deliberately not a lights-out black box by default. Humans define intent, architecture, and acceptance criteria; agents implement inside traceable worktrees; reviewers accept, reject, or merge with an audit trail.

The goal is not more prompt text. The goal is a durable operating system for agentic coding where the repository remains the source of truth.

Is It For You?

Use Spec Kitty when:

• AI coding sessions are losing requirements, decisions, or acceptance criteria.
• You want specs, plans, tasks, reviews, and merge state stored in Git.
• Multiple agents or developers need clear work-package boundaries.
• You are running parallel Claude Code, Codex, Cursor, Copilot, Gemini, or Windsurf work and need git worktree isolation.
• You are moving from vibe coding to a repeatable spec-driven development workflow.
• You want a local-first workflow where the repository stays the source of truth.

It is probably overkill for one-off edits, tiny scripts, or teams that do not use Git.

What It Provides

Need	Spec Kitty provides
Start from intent	Guided specify, plan, and tasks workflows
Keep agents aligned	Repository-native mission artifacts under kitty-specs/
Split implementation	Work packages with lifecycle lanes such as planned, in_progress, for_review, approved, and done
Run agents in parallel	Isolated git worktrees under .worktrees/
Keep quality visible	Review, accept, merge, and retrospective gates
See progress	Lane-by-lane mission status with spec-kitty agent tasks status (--json for tools)
Integrate agents	Slash commands or skills for Claude Code, Codex, Cursor, Gemini, Copilot, Windsurf, OpenCode, and more
Learn from missions	Every completed mission generates a retrospective by default. Tune via .kittify/config.yaml#retrospective or charter; see how-to.

Common Use Cases

• Replace ad hoc vibe coding with spec-driven development.
• Turn GitHub issues, product requirements, or bug reports into executable work packages.
• Coordinate multiple AI coding agents without losing context between sessions.
• Keep architecture decisions, constraints, and acceptance criteria close to the code.
• Build a governed software factory that can scale toward more autonomy without hiding review, test, or merge decisions.

Governance layer

Spec Kitty keeps runtime governance in the repo instead of treating it as agent-only prompt text. The trail model in docs/architecture/trail-model.md describes how spec-kitty dispatch "<request>" maps operator intent to runtime behavior, while docs/architecture/host-surface-parity.md tracks parity across CLI, slash-command, and hosted surfaces.

The primary standalone governance command is:

• spec-kitty dispatch "<request>" - loads governance context, opens an Op record, and returns the context the agent must use before doing the work

Quick Start

Install the CLI:

pipx install spec-kitty-cli

pipx is the preferred installer for the CLI because it keeps Spec Kitty in its own virtual environment and avoids the externally-managed-environment errors common on modern Linux distributions.

Other supported install methods:

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

Create or initialize a project:

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

Replace claude with your agent key when needed. Common choices include codex, cursor, gemini, copilot, opencode, qwen, windsurf, kiro, vibe, pi, letta, antigravity, and llxprt. See Supported Agents for the current list.

Open your AI coding agent in the project and run the core workflow:

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

Agents that load Spec Kitty as Agent Skills (Codex, Vibe, Pi, Letta and Antigravity) run the same commands as skills named spec-kitty-<command>: for example $spec-kitty-specify in Codex, /skill:spec-kitty-specify in Pi and /spec-kitty-specify in the others.

Then let the runtime choose the next action until the mission is ready:

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

Review, accept, merge, and close the loop:

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

After merge, run /spec-kitty-mission-review. The mission's retrospective.yaml is authored during the runtime terminus (HiC prompt or autonomous facilitator), not by merge. Once it exists, use spec-kitty retrospect summary for the cross-mission view and spec-kitty agent retrospect synthesize --mission <mission-slug> to apply any staged proposals (dry-run by default — pass --apply to mutate).

For the full walkthrough, see Your First Mission.

Everyday Commands

Command	Purpose
spec-kitty init . --ai <agent>	Add Spec Kitty to the current repo
spec-kitty verify-setup	Check local installation and project wiring
spec-kitty agent tasks status	Show work package status for a mission
spec-kitty next --agent <agent> --mission <slug>	Ask Spec Kitty what the agent should do next
spec-kitty upgrade	Update an existing project after upgrading the CLI
spec-kitty --help	Show available commands

Documentation

Start here:

• Getting Started
• Your First Mission
• Orchestrator Quickstart
• CLI Command Reference
• Slash Commands
• Supported Agents
• Install and Upgrade

Deeper topics:

• Spec-Driven Development
• Mission System
• Git Worktrees
• Multi-Agent Orchestration
• External Orchestrator Runbook

Hosted auth, sync, and tracker flows were removed from open-core (#6164; the code is preserved in spec-kitty/event-router). For background, see Launch-Readiness Behavior.

FAQ

Is Spec Kitty for dark software factories?

Spec Kitty can be used as part of a dark software factory or autonomous coding pipeline, but its default model is governed and human-in-loop. It keeps specs, work packages, agent actions, review decisions, and merge state visible in the repository.

Which AI coding agents does Spec Kitty support?

Spec Kitty supports common AI coding agents and coding harnesses including Claude Code, Codex, Cursor, Gemini, GitHub Copilot, OpenCode, Qwen, Windsurf, Kiro, Vibe, Pi, Letta, and Antigravity. See Supported Agents.

How is Spec Kitty different from prompt templates or Spec Kit?

Spec Kitty is inspired by spec-driven development workflows, but adds repo-native mission state, work-package lanes, git worktree isolation, governance commands, and an explicit next -> review -> accept -> merge runtime loop.

Does Spec Kitty require a SaaS service?

No. Spec Kitty is local-first and stores its core artifacts in your repo. Hosted tracker and sync integrations are optional.

Development

git clone https://github.com/spec-kitty/spec-kitty.git
cd spec-kitty
pip install -e ".[test]"

When testing templates from a source checkout:

export SPEC_KITTY_TEMPLATE_ROOT="$(pwd)"
spec-kitty init my-project --ai claude

See the Contributing guide for contribution guidelines.

Support

• Open a GitHub issue for bugs, feature requests, or questions.
• See the changelog for release notes.
• See CONTRIBUTORS.md and the GitHub contributors graph for contributor credits.

License

Spec Kitty is released under the MIT License.