Spec-Driven Development is not Waterfall in a Hoodie
Andrea Laforgia calls spec-driven development “Waterfall in a Hoodie”. He is right about big batches and approval gates, but the problem is batch size, not the spec. Here is how the AI Unified Process works in thin slices, one use case at a time.
Specs, Guidelines, and Sensors in Spec-Driven Development
Specs define the what, guidelines define the how. But a guideline without an automatic check is only a wish. This article shows how to give every guideline and ADR a sensor, using the decision “jOOQ, not JPA” with Maven Enforcer and ArchUnit as the example.
Use Cases and Architecture: The Unified Process in the Age of AI
The Unified Process has always been use-case driven and architecture-centric. With AI agents, these two principles matter more than ever. Here is how the AI Unified Process supports both with AI.
Not a Factory: Why AI Development Is a Workshop
AI did not turn software development into a factory. It is a workshop: the agent is the tool, the spec is the plan, and the software engineer is the craftsman who owns the harness. And like in Santa’s workshop, a vague wish list gives you socks instead of a bike.
Passkeys in Vaadin Flow with Spring Security
How to add passkey sign-in to a Vaadin Flow application with Spring Security’s built-in WebAuthn support, including the two Vaadin-specific details you need to know.
Five Agents, Fifty Decisions
Adam Tornhill shows that five parallel agents mean about 50 tasks for the developer. The way out is not more agents, but making the hidden design decisions once, in the specification. Here is how the AI Unified Process does this.
Good Developers Are Not Wary. They Are Asking the Right Questions.
Rachel Laycock asks why good developers are wary of agentic development. Her questions are the right ones, but many of them already have answers. Here is how Spec-Driven Development and the AI Unified Process answer them, and where Rachel is right.
OpenSpec and the Missing Flow from the User’s Perspective
OpenSpec describes very precisely what a system must do. But nowhere does it show, in one piece, how a user moves through the application step by step. This is not a bug in the tool. It is a deliberate decision of the format, and it has consequences. I migrated the Spring PetClinic to OpenSpec. The [...]
Who Owns the Test Oracle?
More than 12,000 unit tests, all green. And still the bugs show up at the release boundary. This is what Paul Stack reports from Swamp, where agents write every line of code. In two recent posts, Your Agent-Written Tests Aren’t Real Tests and Why Agent-Written Code Needs Product Verification, he explains why. I agree with [...]
Reverse Engineering Existing Code into Specs
Spec-Driven Development works great when you start a new project. You write the specs first, and the AI agent implements them. But most of us don’t work on new projects. We work on systems that are five, ten or twenty years old. Often there are no specs at all, or the specs are outdated.
From Prompt Template to Skill: The IREB AI4RE Prompt Guide
“Write user stories for an ordering portal.” Most requirements prompts look like this. The answer looks plausible, reads well, and nobody can say whether it is good. The next person asks the same thing in different words and gets a different answer. That is not requirements engineering. That is a chat.
Hooks That Stop the Agent From Claiming “Done”
An AI agent that says “done” is making a claim. Claude Code hooks let the session ask for proof: a Stop hook that wants a fresh, green test report, and a status guard that wants a coverage audit before a use case is marked as done.


