Appofficial publisher

spec-kit (GitHub)

App : A product you use directly

Spec-driven development toolkit: constitution, spec, plan, tasks, then implementation.

The most portable discipline across agents to stop coding by gut feeling.

GitHub · 142,754 ★

Overall score

9.2 /10

Security: nothing to flag

Utility

8.5

Reliability

10.0

Security

No caveat on record

Ease

5 min · Comfortable

Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.

What it does, concretely

When a feature is handed off to an AI coding agent, the agent tends to start coding immediately without fully understanding what is expected. Spec Kit solves this by enforcing a specification step before any code is written: you first define what you want to build, then how, then break it into tasks, and only then implement.

In practice, after installing via the specify CLI, you initialize a project and chain slash commands inside your coding agent: /speckit.constitution to establish project principles, /speckit.specify to describe the need, /speckit.plan to choose the architecture, /speckit.tasks to break things into actionable tasks, and /speckit.implement to execute. A /speckit.converge command then checks that the implementation aligns with the spec.

Spec Kit is aimed at developers who use AI coding agents such as GitHub Copilot, Cursor, Claude Code, or other compatible tools, and who want to structure their process to get more predictable and documented results. Optional extensions also cover bug fixing and idea assessment before commitment.

The tool is less suited to people who simply want to generate code quickly without going through a structured thinking phase, or to those who do not work with a compatible coding agent.

What it offers

·Six-step guided workflow (constitution, spec, plan, tasks, implementation, convergence) to structure any agent-driven development.
·The /speckit.converge command automatically checks that the produced code aligns with the defined specification, plan, and tasks.
·Optional bug-fixing extension with an assess, fix, test sequence to keep each patch documented and scoped.
·Optional idea assessment extension that produces a go / needs-clarification / kill decision before any development commitment.
·Compatibility with more than 30 coding agents, including GitHub Copilot, Cursor, Claude Code, and the CLI.
·Extension, preset, and bundle system to adapt the kit to a specific role, team, or context.
·The specify CLI with built-in update commands (specify self upgrade) and a non-interactive mode for CI environments.

Our tips

Repeat the /speckit.implement and /speckit.converge steps in a loop until the converge command reports a "Converged" status: that is the signal that the implementation matches the spec.
In a CI environment or an automated agent without a keyboard, add --non-interactive to the specify init command to prevent any blocking on an interactive picker.
The constitution (/speckit.constitution) is a one-time step per project: take the time to properly define quality, testing, and performance standards before starting to specify features.
If you assess an idea with the assess extension and get a "go" decision, you can move directly to /speckit.specify to continue with the development workflow.
Use specify self check to see whether a newer version is available without changing anything, then specify self upgrade to update in place.

Watch out

Installation requires the uv tool (Python package manager): if you do not already have it, you need to install it separately before using Spec Kit.

In practice

Price: Free
Commercial use: Commercial use: yes
Available on: cli, claude-code, copilot, cursor
Licence: MIT

Security and data

Rating: Security: nothing to flag
Your data: Data stays local
Maintenance: Active
Last release: 22 October 2026

Links and repository checked on 25 October 2026 · added on 1 September 2025

How to install spec-kit (GitHub)

1. Install uv if not already present (see https://docs.astral.sh/uv/). 2. Install the CLI: uv tool install specify-cli (also available from PyPI with the same command). 3. Initialize a project specifying your agent: specify init my-project --integration copilot. 4. Move into the project directory: cd my-project. 5. Launch your coding agent in that directory and start with /speckit.constitution. 6. Check for updates at any time with specify self check, with no risk of modification.

uv tool install specify-cli

In the same category

Whole category Frame (methods, guardrails) →