Spec Kit

Spec Kit is an open-source toolkit from GitHub that introduces a structured, specification-driven development workflow for AI coding agents. Rather than relying on ad-hoc prompts and vibe coding, Spec Kit provides a multi-phase process where developers first define what they want to build through detailed specifications, then create technical implementation plans, break those plans into actionable tasks, and finally let the AI agent execute the implementation.

The toolkit centers around the Specify CLI, a command-line tool that bootstraps projects with the necessary scaffolding for spec-driven development. Once initialized, developers interact with their AI coding agent through a series of slash commands that guide the workflow. The /speckit.constitution command establishes governing principles and development guidelines. The /speckit.specify command captures functional requirements and user stories. The /speckit.plan command creates a technical implementation plan based on the chosen tech stack. The /speckit.tasks command generates an ordered task breakdown with dependency management and parallel execution markers. Finally, /speckit.implement executes all tasks according to the plan.

Spec Kit is agent-agnostic by design, supporting a wide range of AI coding environments including GitHub Copilot, Claude Code, Gemini CLI, Cursor, Windsurf, Codex CLI, Amazon Q Developer CLI, Amp, Roo Code, Kilo Code, opencode, Qwen Code, IBM Bob, and others. Because it works by scaffolding files and prompts within the project repository, it integrates into existing workflows without vendor lock-in. The CLI downloads templates from the GitHub repository and sets up agent-specific configurations automatically.

The development methodology supports three primary phases. Greenfield development starts from scratch with high-level requirements and generates production-ready applications. Creative exploration enables parallel implementations across different technology stacks and architectures. Brownfield modernization handles iterative feature addition and legacy system upgrades. Additional optional commands like /speckit.clarify help resolve ambiguities before planning, /speckit.analyze performs cross-artifact consistency checks, and /speckit.checklist generates quality validation checklists.

Spec Kit requires Python 3.11 or later and uses uv for package management. It runs on Linux, macOS, and Windows, with both shell and PowerShell script variants available. The toolkit supports Git-based feature branching workflows where each feature gets its own specification directory, and also provides an environment variable override for non-Git repositories. Projects are organized with a .specify directory containing memory, scripts, specs, and templates subdirectories.

The project has gained significant traction in the developer community with over 74,000 GitHub stars and more than 6,000 forks. It also supports an extension system that allows community-contributed workflow additions and customizations beyond the core spec-driven development commands.