Who it is for
For developers who want to plan, build, review, test, and ship with agents—without losing control of their codebase.
Agentic AI Coding / Intermediate
The Agentic Workflow
The repeatable loop: plan, build, review, test, and ship with agents without losing control of your codebase.
The problem
It starts with the situation you're actually in.
Replace ad-hoc chatting with agents with a repeatable five-phase loop—plan, build, review, test, ship—and learn why self-orchestrating agents stall, skip steps, or create circular dependencies on real codebases.
The real test isn't whether you can run the loop once, with full attention and the guide in front of you. It's whether you run it on an ordinary Tuesday when a ticket lands, the deadline is close, and the fast move is to paste a vague request into an agent and hope.
Outcomes
What you'll be able to do.
Bound the problem before you delegate it
You write a clean spec that tells an agent what "done" looks like. Hand over a well-formed brief for anything bigger than a single function, and the agent reads the codebase, forms a view of the architecture, and builds—instead of starting, hitting an early decision that blocks a later one, and backtracking.
Prompt for the smallest diff that satisfies the task
With a bounded, ordered plan in hand, you give the first task to an agent and prompt for the smallest diff that does the job—then watch it work.
When the Agent Says "Fixed"
You hand an agent a failing build. Two minutes later it returns a diff and a confident note: the authentication bug is resolved, the test should pass now. You know how to confirm that before you trust it.
The Last Gate Is the One You Own
A passing test suite tells you the code does what you asked—not that it's ready to ship. You own the final gate that closes the loop and decides what goes out.
Inside the book
A closer look at the work inside.
Map your current workflow
Before you adopt the loop, find where you actually lose control today. Take one recent task you handed to an agent—ideally one that went sideways—and walk it through an honest audit, starting with the plan: did you write down what "done" meant before you began?
Practice
Take one real task from your backlog—a small feature or a live bug. Write it first as you normally would, in a single sentence, then expand it into the full template. Every decision you're forced to make in the expansion is one the one-sentence version would have handed silently to the agent—and the rework you just avoided.
Is this spec ready to hand off?
Run this before the agent ever sees the brief. If you cannot answer cleanly, you are not ready to leave the first gate.
- Could a competent stranger build this without asking me a question?
- Is "done" something I can observe or test, not just describe?
- Have I named where the change lives—and where it must not?
- Did I state the invariants that must survive the change?
- Are the failure and empty-input paths defined, not assumed?
- Are my own open questions resolved, not deferred to the agent?
Worked example: turning a bug report into a brief
Start with a real, vague report of the kind that lands in a tracker: "Users say the export button sometimes does nothing." It's unactionable. "Sometimes" and "nothing" are exactly the gaps an agent will fill on its own.
Build your task graph
Take a task you've already specified and run it through the sequence. Use a prompt to have an agent propose a decomposition you then correct—drafting is delegable; the ordering judgment is not. The agent reads the spec, writes no code, and proposes a breakdown into sub-tasks.
Map your current workflowInside the book
Visual preview
A diagram you can keep open while you work.
Table of contents
8 chapters, built to be read in order.
01
The Loop That Keeps You in Charge
02
Writing the Specification an Agent Can Execute
03
Planning and Decomposing the Work
04
Building with Agents Without Losing the Thread
05
Reviewing Agent Output Like a Senior Engineer
06
Testing as the Agent's Accountability Layer
07
Shipping Safely and Keeping the Codebase Yours
08
Your Repeatable Practice
152
Pages
13,850
Words
42
Exercises, checklists & tools
8
Chapters
Formats
Three formats. One purchase.
EPUB, PDF, and HTML are included so the book can work on an e-reader, as a designed copy, or as a searchable desk reference.
EPUB
For e-readers and reading apps.
The designed edition with diagrams and layouts intact.
HTML
Searchable, copy-pasteable, and practical as a reference.
Complete guide
The Agentic Workflow
$9.99
Intermediate guide pricing
- ✓EPUB + PDF + HTML included in one purchase
- ✓8 chapters from the complete guide
- ✓DRM-free files for your own devices
- ✓7-day refund review for duplicate purchases, access issues, wrong files, or materially defective downloads
Secure checkout / instant download / tax handled at checkout
Before you buy
Questions, answered.
Does this include the full book?
Yes. You get the complete edition, including the chapter sequence and internal materials described on this page.
Which formats are included?
EPUB, PDF, and HTML are included so you can read on an e-reader, keep a designed copy, or use the searchable browser version.
What is the refund policy?
Because this is an instant digital download, broad change-of-mind refunds are not offered after the files have been accessed. Refund requests are reviewed within 7 days for duplicate purchases, accidental purchases before access, access failures we cannot fix, wrong files, corrupted files, or pages that materially misdescribe the book.