An open-source AI coding agent that runs in your terminal, with companion desktop, web, editor, and chat interfaces.
github/spec-kit
An open source toolkit that gives AI coding agents structured spec-driven processes, templates, and documented outcomes for building features, fixing bugs, or assessing ideas.
- Stars
- 142,311
- Forks
- 12,743
- Open issues
- 177
- Last commit
- 6 hours ago
What is it?
Spec Kit helps developers work with AI coding agents through explicit specifications rather than ad hoc prompts. Its core is Spec-Driven Development: you define what to build and why, then the agent turns that specification into a plan, tasks, and working code, with commands for implementing and converging the result. Bug fixing and idea assessment are offered as separate entry points, the former adding assess, fix, and test stages that end in a verdict of verified, partial, or failed.
The toolkit is built from composable mechanisms. Integrations connect it to different coding agents, while extensions, presets, workflows, and bundles let teams add commands, adapt templates, automate steps, or package a setup for a role. Catalogs support discovering community components and hosting private ones, and offline or self-hosted sources are supported. Installation uses Python 3.11+ and the uv tool on Linux, macOS, or Windows.
Key features
- 01Spec-Driven Development workflow from constitution through plan, tasks, implementation, and convergence.
- 02Bug-fixing extension that assesses cause, applies a scoped fix, and records verification.
- 03Idea-assessment extension that produces a go, needs-clarification, or kill decision.
- 04Integrations for switching between coding agents, with GitHub Copilot as the example.
- 05Extensions, presets, workflows, and bundles for customizing or sharing processes.
- 06Community and private catalogs for discovering and hosting components.
Use cases
- Turning a feature idea into a specification, plan, and implementation with an AI coding agent.
- Diagnosing and verifying a bug fix with recorded reports before merging.
- Deciding whether a product idea deserves engineering investment before writing code.
- Packaging approved templates and standards for an engineering team to reuse across projects.
Quick start
- 1
Install the Specify CLI with uv
uv tool install specify-cli - 2
Create a project for GitHub Copilot
specify init my-project --integration copilot - 3
Enter the project directory, then launch your coding agent there and run the spec-driven skills in its chat
cd my-project
Things to know
- Requires Python 3.11+ and the uv tool to be installed first.
- Workflows depend on a supported AI coding agent and its chat interface.
Topics
FAQ
- What is spec-kit?
- An open source toolkit that gives AI coding agents structured spec-driven processes, templates, and documented outcomes for building features, fixing bugs, or assessing ideas. Spec Kit helps developers work with AI coding agents through explicit specifications rather than ad hoc prompts. Its core is Spec-Driven Development: you define what to build and why, then the agent turns that specification into a plan, tasks, and working code, with commands for implementing and converging the result. Bug fixing and idea assessment are offered as separate entry points, the former adding assess, fix, and test stages that end in a verdict of verified, partial, or failed.
- Is spec-kit free to use? What license does it have?
- spec-kit is open source under the MIT license. Check the license terms before using it in commercial products.
- Is spec-kit still maintained?
- Actively maintained. The last commit to spec-kit was on Oct 21, 2026.
- What can I use spec-kit for?
- Turning a feature idea into a specification, plan, and implementation with an AI coding agent. Diagnosing and verifying a bug fix with recorded reports before merging. Deciding whether a product idea deserves engineering investment before writing code. Packaging approved templates and standards for an engineering team to reuse across projects.
- How do I install spec-kit?
- Install the Specify CLI with uv: uv tool install specify-cli Create a project for GitHub Copilot: specify init my-project --integration copilot Enter the project directory, then launch your coding agent there and run the spec-driven skills in its chat: cd my-project
- Where can I get spec-kit?
- The source code and installation instructions are on GitHub: https://github.com/github/spec-kit
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source browser automation that uses LLMs and computer vision to run web workflows on any site, extending Playwright with AI commands.
A Neovim plugin that brings Cursor-style AI chat, code suggestions and one-click diff application to your editor.
An open-source LLM application devops platform for enterprises, covering workflows, RAG, agents, model management, evaluation and fine-tuning.
TypeScript framework for building AI agents with memory, tools, MCP, RAG, voice and workflows, plus the VoltOps console for observability.
Open-source AI coding agent for VS Code that works with your own models, including hosted APIs, accounts, and local Ollama or llama.cpp.