Map the real work
Identify your codebase, tools, constraints, and the decisions only a person should make.
A practical guide to AI workflows
Automate your work and free your entire week.
A clear method for turning repeated development work into a process an agent can follow and you can verify.
By Raphael Bernardo · English ebook
Choose your preferred store. Local prices and availability may differ.

Inside the book
Identify your codebase, tools, constraints, and the decisions only a person should make.
Write reusable skills—short procedures an AI agent can load when they apply.
Use observable acceptance checks instead of treating a confident answer as proof.
A real starting point
A pull request is a proposed code change submitted for review. Suppose you regularly write its description from a ticket and a set of edits. Instead of asking an agent to invent a polished summary, give it the ticket, the actual file changes, and the test output you really ran.
Then ask it to draft only claims supported by that evidence. You remain responsible for checking the result before sharing it. The same pattern works for release notes, review checklists, and other repeated tasks.
Who this is for
If you use Git and a terminal, this book gives you small examples you can adapt to a real project. It covers context mapping, reusable instructions, connecting an agent to tools, and checking results. You do not need to adopt a single vendor's full stack to use the ideas.
Live integrations require their own accounts, permissions, and setup. The examples are designed to start small, with local checks before you connect a production service.
Quick answers
The edition linked here is in English.
No. The method is about mapping work, writing reusable procedures, and verifying results. Tool setup varies by provider.
Choose the store where you already read ebooks. The stores control their own pricing, payment, and regional availability.
Start with one task
Find the ebook in the store you prefer.