Spec-Driven Development Toolkit from GitHub
6 days ago
- #Software Development Methodology
- #AI-assisted coding
- #Spec-Driven Development
- Spec-Driven Development (SDD) shifts focus from code-first to specification-first, making specs executable and central to development.
- Initialization involves setting up a project with an AI coding agent (Claude Code, GitHub Copilot, Gemini CLI) using the `specify init` command.
- Key commands include `/specify` for defining project requirements, `/plan` for tech stack details, and `/tasks` for actionable implementation steps.
- SDD emphasizes intent-driven development, rich specification creation, multi-step refinement, and leveraging AI for specification interpretation.
- Development phases include 0-to-1 (greenfield), creative exploration (parallel implementations), and iterative enhancement (brownfield).
- Experimental goals include validating SDD's applicability across tech stacks, supporting enterprise constraints, and enabling diverse development approaches.
- Prerequisites include Linux/macOS/WSL2, AI coding agents, Python 3.11+, Git, and package management tools like `uv`.
- The process involves detailed specification drafting, tech stack planning, implementation, and troubleshooting with AI assistance.
- Troubleshooting tips include resolving Git authentication issues and handling over-engineered components by adhering to project constitutions.