Spec Kit — spec-driven development with AI agents

★ 8.0 · vibe-coding

Spec Kit is an open-source toolkit of processes, templates and skills for AI coding agents, invoked as /speckit-* commands in the agent's chat, plus a specify CLI for setup. Reach for it when you want to pin down what and why before writing code; when your agent keeps drifting without a clear plan; when a bug should be fixed by an assessed cause rather than a symptom; or when you need to decide whether an idea deserves investment. There are three independent processes: spec-driven development (constitution, then specify, plan, tasks, implement, converge), bug fixing (assess, fix, test, ending in a verified, partial or failed verdict) and idea assessment (intake, research, define, shape, decide, ending in go, needs-clarification or kill). Spec-driven development ships in core, while bug and assess are installed as extensions with specify extension add. Install with uv tool install specify-cli, then specify init my-project --integration copilot, replacing copilot with the integration key of another supported agent. You need Python 3.11 or newer, uv and a supported agent on Linux, macOS or Windows. These are skills in the agent's chat, not terminal commands. It does not replace a test framework or CI.