The definitive guide to Spec-Driven Development: build software with AI agents, from method to production code. Book in EN + pt-BR with a ready-to-use SDD kit for Claude Code.
This repository also configures its own agents. See what spec-driven-development-book tells them →
felipefontoura/spec-driven-development-book
Agent Claude Code
Produces requirements.md and design.md from an approved idea or feature request. Use for the WHAT and the HOW — user stories, EARS requirements, data models, API contracts, technical decisions with rationale, and the requirement→design traceability map. Never writes implementation code.
felipefontoura/spec-driven-development-book
Agent Claude Code
Implements one approved task from tasks.md — code plus tests — after checking that .status reads tasks:approved. Follows the spec exactly, stops on ambiguity instead of guessing, and reports verification evidence. Use for the EXEC phase, one task at a time.
felipefontoura/spec-driven-development-book
Agent Claude Code
Reviews an implementation against requirements.md, design.md, tasks.md, and the produced code in a fresh context. Reports coverage gaps and spec violations with verification evidence — not stylistic opinions. Use before treating a feature as done. Reviews only; does not implement fixes.