Spec Kit

Overview

Spec Kit is an open source toolkit that gives AI coding agents structured processes, reusable templates, and documented outcomes. Its core Spec-Driven Development flow carries a specification through planning, implementation, and convergence, while optional extensions support bug fixing and idea assessment.

Key Features

• Core Spec-Driven Development process: Specify, Plan, Tasks, Implement, Converge, with each phase producing a Markdown artifact that feeds the next.
• 38 coding-agent integrations, including Copilot, Gemini, Codex, Claude, Zed, and Kiro, with a generic option for other tools and single-command switching.
• Optional bug-fixing extension that assesses the cause, applies a scoped fix, and records verification with a verified, partial, or failed verdict.
• Optional idea-assessment extension that gathers evidence and produces a go, needs-clarification, or kill decision.
• Customization through extensions, presets, workflows, bundles, project-local overrides, and community or self-hosted catalogs.
• Offline and air-gapped operation with local sources and self-hosted catalogs on Windows, macOS, and Linux.

Use Cases

• Build a feature or application from a specification that is carried through planning, implementation, and convergence.
• Diagnose and repair broken behavior with an assessed cause, a scoped fix, and recorded verification.
• Decide whether an idea deserves investment using an evidence-backed go, clarify, or stop decision.
• Adopt or create a different process for team standards, organizational governance, or non-software workflows.

Getting Started

• Ensure Python 3.11+, uv, and a supported AI coding agent are available. Install Spec Kit with uv tool install specify-cli, create a project with specify init my-project --integration copilot, and launch your coding agent in the project directory. For bug fixing, run specify

Deployment & Requirements

• Requires Python 3.11+, uv, and a supported AI coding agent on Linux, macOS, or Windows. Spec Kit works offline, behind firewalls, and in enterprise or air-gapped setups using local sources and self-hosted catalogs.

Before You Adopt

• License: MIT. Review its terms before using, modifying, or distributing the project.
• Adoption requires Python 3.11 or newer.
• A supported AI coding agent is required on Linux, macOS, or Windows.
• Bug fixing and idea assessment are bundled, opt-in extensions rather than mandatory core phases.
• Community catalog listings are not audits or endorsements, so review source code before installing or running it.