A0 · chapter overview
Why an agile agent workflow
The foundations, in three questions. Why a pragmatic, story-led, agent-driven workflow beats both vibe coding and big-bang specs. What we are building — the framework the course runs on, in two layers, four artifacts, and one loop. And who does the work — the two roles that drive every rung. Select a question to see its module.
The chapter · why, what, who
In this chapter
Three modules carry the foundations. A0.2 — What we are building is written now; it lays out the framework every later chapter applies. A0.1 — Why it works and A0.3 — Who does the work are on the way. Each module holds three subpages.
Why it works
The case: the two ways software built with AI fails, and why thin, provable slices beat both.
- A0.1.1 The two failure modes
- A0.1.2 Pragmatic Programming, for agents
- A0.1.3 Correct by definition
What we are building
The framework the course runs on: its structure, its vocabulary, and how it moves.
Who does the work
The pairing: the human who decides and accepts, the agent who specifies and builds, and the rhythm between them.
- A0.3.1 The Operator
- A0.3.2 The Author
- A0.3.3 The pairing in practice
Start with the built module, A0.2 — What we are building: read A0.2.1 · The two-layer model first. The next chapter, A1 · Decomposition, turns this framework into a backlog of user stories.
References
Primary sources for the foundations, and where the chapter connects.
Sources
- Hunt, A. & Thomas, D. — The Pragmatic Programmer — tracer bullets and tight feedback.
- Beck, K. — Extreme Programming Explained — small batches, short loops.
- Adzic, G. — Specification by Example — the spec as one living source of truth.