← Back to archive

Repo of the Day

github/spec-kit

Published: Sep 5, 2025

Open repository ↗

💫 Toolkit to help you get started with Spec-Driven Development

Summary

Spec Kit is an open source Python toolkit (MIT licensed) that gives AI coding agents structured processes, reusable templates, and documented outcomes. It bundles three independent entry points — Spec-Driven Development, bug fixing, and idea assessment — so engineers can pick one without committing to the others.

What it is useful for

Spec Kit is useful when you want an AI coding agent to follow a defined process rather than improvise. The core SDD flow turns a feature request into a specification, a technical plan, and tasks, then checks the implementation against those artifacts with a converge step that reports "Converged" when the work matches the spec. The opt-in bug fixing extension separates diagnosis, repair, and verification, producing a verified, partial, or failed verdict stored under .specify/bugs/<slug>/, and the README explicitly states that "missing verification is not a successful fix." The opt-in idea assessment extension produces a go, needs-clarification, or kill decision stored under .specify/assessments/<slug>/, so a team can stop work on a weak idea with a documented reason instead of just shelving it. The toolkit also supports customization through extensions, presets, workflows, and bundles for role-based setups.

How engineers can use it

The README documents the following setup. You need Python 3.11+, the uv tool, and a supported AI coding agent on Linux, macOS, or Windows. For a new project:

uv tool install specify-cli
specify init my-project --integration copilot
cd my-project

Skills are invoked in the agent's chat, not the terminal. The SDD flow is /speckit-constitution once per project, then per feature: /speckit-specify, /speckit-plan, /speckit-tasks, /speckit-implement, /speckit-converge, repeating the last two until convergence. A documented example is a photo organizer built with Vite and vanilla JavaScript, with metadata stored in SQLite.

Bug fixing needs specify extension add bug, then /speckit-bug-assess, /speckit-bug-fix, and /speckit-bug-test with a slug. Idea assessment needs specify extension add assess, then /speckit-assess-intake, /speckit-assess-research, /speckit-assess-define, /speckit-assess-shape, and /speckit-assess-decide. The README notes that non-Copilot agents and other modes may use different invocation syntax, so the integrations reference should be checked first. For pinned releases, CI setup, or adding Spec Kit to an existing codebase, the README points to the installation and existing-project guides on the documentation site rather than covering those paths inline.