New Second edition · September 2026

The Agentic Crew

Engineering in the age of AI agents

A field guide for software engineers learning to work with agents instead of against them — context, guardrails, tests, conventions, and the judgement that ties them together. Written by an engineer who ships with agents every day. Free, forever.

  • 19chapters
  • 2appendices
  • 4languages
  • €0always free
The Agentic Crew, second edition — book cover with a gold ship's helm on midnight blue
Second edition

Six months is a long time at sea.

The first edition shipped in March 2026. Since then the tools converged, cloud agents started sending pull requests, and prompt injection went from a curiosity to a string of real incidents. This edition catches up — and fixes what I got wrong.

i New chapter

The Agent Attack Surface

Prompt injection, the lethal trifecta, poisoned tools, and slopsquatting — and the layered defences that keep a stranger's text from steering your agent.

ii New chapter

Building Your Own Agents

From driving a coding agent to shipping your own: owning the loop and the context, humans as a tool call, durable state, and evals.

iii Rewritten

Context engineering & the new toolchain

Compaction, context rot, sub-agents, AGENTS.md, skills, hooks, cloud agents that return PRs — and a rewritten chapter on models.

iv New appendix

The State of the Tools

The perishable details — models, prices, products — moved to one dated appendix, so the chapters can age gracefully.

Why this book

The ground is shifting.

It started with porting the Claude Agent SDK from Python to Go — using Claude Code itself. Over a hundred prompts, and one realisation: most of them were wasted, because I didn't yet know how to direct the work. Not because the model wasn't smart enough. Because I wasn't.

  • i

    The craft is molting, not dying

    The keystrokes and boilerplate fall away. Systems thinking, taste, and judgement become the whole game.

  • ii

    Nobody handed us a playbook

    The tools arrived faster than the practices. This is the honest engineering guide — no hype, no doom, real scar tissue.

  • iii

    The boring things win

    Context, guardrails, tests, conventions. Tools change every quarter. Those compound over a career.

Watercolour of a robot crew working the sails of a wooden ship
Plate I — The crew takes the sheets; the captain keeps the helm.
The craft isn't dying. It's molting.

The outer shell — the keystrokes, the syntax, the boilerplate — that part is falling away. But the animal underneath? The part that knows what to build and why, that smells a bad abstraction from three files away? That part is more alive than ever.

From the Foreword
What's inside

The voyage, chapter by chapter.

Nineteen chapters in five parts, from what an agent really is to how whole teams work with them — plus two appendices.

Or watch it: a two-minute film per chapter

I

Setting Sail

What changed, and what an agent really is

  1. 01

    Introduction

    The fundamental loop of software engineering is breaking. What it means to go from writing software to engineering it — with a crew of agents at your side.

    1:46 film
  2. 02

    What Is an Agent?

    Model, tools, loop, and harness. What makes something truly agentic, how tool calling actually works, how agents fail, and the right mental model for working with them.

    2:36 film
II

Rigging the Ship

The durable foundations: context, guardrails, git, tests, conventions, memory

  1. 03

    Context

    The most important skill in agentic engineering is context engineering. Curating what agents see, compaction, sub-agents for isolation, and why bigger windows didn't remove the tax.

    2:33 film
  2. 04

    Guardrails, Trust, and Sandboxes

    The trust gradient, real permission configs, hooks the model can't talk its way around, sandboxing, and the ethics of delegation.

    2:35 film
  3. 05

    Git as Agent Infrastructure

    Small commits, descriptive messages, branch-per-task, worktrees for parallel agents, and reviewing at the volume agents produce.

    2:30 film
  4. 06

    Testing as the Feedback Loop

    Tests become the agent's eyes. TDD as a superpower, fast test suites as agent infrastructure, and the virtuous cycle of better tests and more autonomous agents.

    2:45 film
  5. 07

    Convention Over Configuration

    Consistent structure, AGENTS.md instruction files, and skills — conventions your agents can execute, not just read.

    2:34 film
  6. 08

    The Ship's Log

    Memory for agentic workflows — capturing knowledge across sessions and agents, separating knowledge from execution state, and keeping memory safe.

    2:28 film
III

Beyond the Harbour

Reach, risk, intent, and the models underneath

  1. 09

    Extending the Agent's Reach

    MCP, remote servers, and connecting agents to databases, monitoring, and project management — without drowning them in tools.

    2:35 film
  2. 10

    The Agent Attack SurfaceNew

    Prompt injection, the lethal trifecta, poisoned tools, slopsquatting, and layered defences that limit the blast radius.

    2:38 film
  3. 11

    Articulating Intent

    From doing to articulating. Plan-first workflows, specs, constraints, task decomposition, voice and visual context, and skills as the modern prompt library.

    2:28 film
  4. 12

    Local, Commercial, and Hybrid Models

    Capability, cost, privacy, and speed. Open-weight models, subscriptions versus API spend, honest ROI, and matching the model to the task.

    2:38 film
IV

Running a Fleet

Many agents, pipelines, and building your own

  1. 13

    Multi-Agent Orchestration

    Sub-agents for context isolation, background agents that return PRs, the handover pattern, and knowing when more agents isn't better.

    2:37 film
  2. 14

    Agents in the Pipeline

    Agents in CI/CD — review bots, the overnight agent turned product, cost control, and untrusted input when nobody's watching.

    2:26 film
  3. 15

    Building Your Own AgentsNew

    From driving a coding agent to shipping your own — owning the loop, the context, human approvals, durable state, and evals.

    2:38 film
V

Hard-Won Lessons

Failure, restraint, teams, and what lasts

  1. 16

    When Agents Get It Wrong

    War stories from the field — the eager refactorer, the hallucinated library, the poisoned ticket — plus a diagnostic playbook for fixing agent failures.

    2:39 film
  2. 17

    When Not to Use Agents

    The overhead tax, architecture decisions, security-critical code, hostile input, legacy codebases, and the craft argument.

    2:30 film
  3. 18

    Agentic Teams

    Measured versus perceived productivity, code review at volume, a path for junior engineers, knowledge distribution, compliance, and hiring.

    2:52 film
  4. 19

    Final Words

    The craft isn't dying — it's shedding its skin. What changed my mind since the first edition, and why the captain still matters.

    2:31 film
Appendix AAgents as Pentesters film Appendix BThe State of the Tools New film
Portrait of Rasmus Bornhøft Schlünsen
The captain

Rasmus Bornhøft Schlünsen

Fifteen-plus years in the software industry — from Django database backends and embedded systems to Rust CLI tools and open-source libraries used by thousands, including django-timescaledb and rana. He works as an AI R&D Engineer at Clovrlabs.

He knows what a production incident feels like at 2am and what it takes to ship real products. These days he spends most of his time building agentic systems — and writing down the playbook he wished he'd had.

Free download

Take it aboard.

No sign-up, no email wall. PDF for reading at a desk, EPUB for e-readers. Download it as often as you like.

Listen free

Hear it on watch.

AI-narrated audiobooks, chapter by chapter: both books, two voices, English and Catalan. It remembers where you stopped.

01
Ready to play

Introduction

The Agentic Crew · English
0:00 8:17
17 chapters · about 4 h 37 min · Narrates the first edition; the second-edition chapters are on the way. Download the whole book (MP3, 147 MB)
Polly
Polly ● Ready to load
Polly

Meet Polly 🦜

Your AI guide to The Agentic Crew.
Runs entirely in your browser — private & offline-capable.

⚠️ Polly is a very small model (0.5B) and may hallucinate. She's trained on the book but take answers with a grain of salt!

~500 MB download · cached for future visits

Runs in any modern browser

How Polly was built →