Spec Kit
github/spec-kitReviewed by AI
What it does
Spec Kit is GitHub's toolkit for working with an AI coding assistant in an orderly way. You install the `specify` command, run it once in a project, and your assistant gains new commands such as /speckit-specify, /speckit-plan, /speckit-tasks and /speckit-implement, each producing a written document you can read before the next step. Extra packs cover bug fixing and deciding whether an idea is worth building, and everything can be adapted with presets and extensions. You need Python 3.11+, the uv installer and an AI coding assistant of your own.
How easy it is to install
C · install 1 or 2 tools · Windows
- Windows C
- macOS C
- Linux C
The same two commands on Windows, macOS and Linux: `uv tool install specify-cli` then `specify init my-project --integration copilot`, which means first installing Python 3.11+ and uv. There is no window and no installer, and from there on you type the commands inside your assistant's chat, so it only makes sense if you already run an AI coding assistant.
Numbers on GitHub
- Stars
- 144,446on Oct 31, 2026
- Downloads
- 943,743222 this week
- Growth
- +1,413stars in 5 days
Stars over time
Facts
- Runs on
- Windows, macOS, Linux
- You need to install
- Python
- To build from source
- Python
- Disk space
- 100 MB–300 MB · estimate
- Accounts or keys
- None
- Cost
- Free · Free and MIT-licensed, published by GitHub. It only supplies the process: the AI coding assistant you drive it with (GitHub Copilot, Claude Code or another) is yours and may well be a paid subscription.
- License
- MIT (Permissive)
- Last update
- Oct 30, 2026
Risk signals (5)
- Attentioncommunity components are not checked by github
README, Capabilities: "Spec Kit includes community catalogs, and organizations can add their own curated sources. Catalog listings are not audits or endorsements, so review community source code before installing or running it." Anything you add with `specify extension add` or `specify bundle search` is a stranger's code.
AI review
- Attentionextensions and workflows can run commands on your machine
The capability table offers to "Automate prompts, commands, gates, and shell tasks" through a workflow, to "React to coding-agent lifecycle events" and to "Run extension commands before or after Spec Kit commands" through hooks. So an installed component can run shell commands around your assistant's work, not only write text.
AI review
- Infoit writes files into your project
`specify init` creates the project scaffolding and the assistant's command files, and the processes leave documents behind in `.specify/`, for example `.specify/bugs/login-crash/` and `.specify/assessments/offline-mode/`. Those folders become part of your repository.
AI review
- Infoyou are still the reviewer
The README repeats for every process: "Run these skills in your agent's chat, not your terminal, one at a time, reviewing each result before continuing." The tool organises the work; the assistant writes the code and the judgement stays with you.
AI review
- Infoyoung project changing almost daily
Created on 2025-08-21 and already at 237 releases, with v1.2.3, v1.2.4 and v1.2.5 between 2026-10-22 and 2026-10-28, and the version in pyproject.toml reading 1.2.6.dev0. Commands and file layout may still move under you.
AI review
Evidence is shown in its original language.
What Radar could not check
- Whether this way of working actually gives better results than asking your assistant directly
- How many AI tokens the full cycle burns, and what that costs on your plan
- Which files each integration writes for your particular assistant, since those details live in the online documentation
- What a community extension, preset or bundle does once installed, which the README itself tells you to review first
- Whether it still works with the version of your coding assistant, given how fast both sides change