Books
Prompt to… — four short books, one argument: an agent gives you the capability, and these teach the judgement the capability assumes you already have. For as long as building software, running an analysis or modelling a business’s finances required real training, the training and the judgement came bundled — you could not get one without absorbing the other. An agent now produces the visible output without the training that used to guarantee the judgement came with it. Each book hands one specific piece of it back.
Software engineering
Prompt to Production
For people who can build software with an agent but are not software engineers. Working code is not the same thing as a system you can run: the vocabulary and the map, and what goes wrong with each part. 17 chapters.
github.com/alpibrusl/prompt-to-production →
Data literacy
Prompt to Evidence
For people who can perform sophisticated analysis with an agent but are not statisticians. How to tell whether a number is load-bearing before you build on it.
github.com/alpibrusl/prompt-to-evidence →
Financial literacy
Prompt to Ledger
For people who can build a financial model with an agent but are not accountants. What the model assumes, where it quietly flatters you, and which figure a reader will act on.
github.com/alpibrusl/prompt-to-ledger →
Decision-making
Prompt to Decision
The layer above the other three. Even a recommendation you have verified is not yet a decision — and that gap turns out to be the series’ real subject, the whole way through.
github.com/alpibrusl/prompt-to-decision →Books as code. The manuscripts are Markdown in Git; the EPUB, PDF and
HTML are build artifacts, not documents anyone edits by hand. They are built with
content-kit, our open-source
books-as-code toolchain, and a continuity linter refuses a build in which a term is used
before it is defined. Each book also carries a cohort curriculum — the same source
rendered into a student handout and a facilitator guide by
cohort-kit — and a
verify-* skill that runs the book’s own checks against an agent’s work.
Manuscripts are CC BY-NC 4.0; the tooling and the curricula are EUPL-1.2.