When code is the consequence.

For decades, the core competence of a software professional has been writing code. SDD —Spec Driven Development— proposes that this competence is shifting. It does not disappear, but it stops being the central act of daily work. What makes you productive now is your ability to clearly specify what you want to build, under what constraints and with what success criteria.

This guide is a complete deep dive into the methodology that Thoughtworks included in its Technology Radar, that GitHub turned into Spec Kit and that AWS materialized as a full IDE with Kiro. It covers the fundamentals, the four-phase methodology, the techniques for writing effective specs, the connection with Scrum and agile frameworks, the tool ecosystem and the mistakes to avoid.

It does not dodge the tensions. It analyzes why SDD looks like waterfall in form but not in substance. It explains when the overhead of specifying does not pay off. And it argues, with reasoning and data, that SDD does not contradict agility but rather completes it in a world where AI agents are part of the team.

24 chapters. A practical case from start to finish. Glossary, verifiable sources and an open license for trainers.

For those who work with AI and want to stop improvising.