For people taking back control
A field guide to AI agents you can stand behind.
An agent can turn a request into a convincing change before you have time to understand its consequences. This book is how you keep that speed—and stay accountable.
Learn to run the machine.
An agent is a machine. It keeps working after you stop touching it, and it builds exactly what you asked for — including the parts you didn't mean. Every powerful tool has offered the same bargain: useless in careless hands, transformative in skilled ones.
For the first time, the limit on what you can build is your intent, not your hands.
That skill is the whole prize. Operate the machine well and the labor moves to it while the judgment stays with you.
The operator’s job.
Where the machine leaves the decision to you.
Know what you’re actually running.
See what the model can and can’t do, and the pieces you assemble around it. The rest of the book builds on this.
Make the agent build what you meant.
Hand over the goals, boundaries, and trade-offs it can’t guess — and keep a long job on course.
Trust results you can verify.
See what the agent sees, catch the drift, and check the work with something other than the agent itself.
Own what ships.
Set the bar before the code does, prove the work cleared it, and keep the final call yours.
Free & open source — nothing to sell. This book, the tools it references, and everything built around them are 100% open source, community-backed projects with no commercial agenda. No paid tiers. No enterprise licenses. Nothing to buy.
-
Book source github.com/
agenticoding/ 109starsagenticoding.github.io -
ChunkHound github.com/
chunkhound/ 1.4kstarschunkhound -
Pi agenticoding extension github.com/
agenticoding/ 51starspi-agenticoding
Next: LLMs Demystified