You already picked an agent back when you set up your workspace. Now you need to understand what it actually is, because the difference between steering it and just hoping comes down to knowing how it works. This chapter gives you that mental model.
1.2.1What an agent actually is
An agent is not a chatbot that answers questions. It reads the files in your project, runs commands, writes and changes code, sees what happened, and goes again, in a loop, until the task is done.
That is the whole shift: a chatbot talks, an agent acts. It works inside your actual project, on your actual files, which is why everything in this book is about directing that action well.
1.2.2What it can and can't do
| It's brilliant at | It cannot |
|---|---|
| writing and refactoring code | know what you actually want |
| wiring up a feature | decide what's worth building |
| reading an error and fixing it | judge when it's good enough to ship |
| explaining a strange file in seconds | own the consequences |
Give it a clear task and it outruns any human at the typing. But it is brilliant at the how and blind to the whether and the why, and that gap is exactly your job.
1.2.3It works from context, not memory
An agent works from context: what is in front of it right now, your message, the files it has open, what it just read. It does not carry a memory of your project between sessions the way a teammate would. Close it, reopen it, and it starts fresh.
So getting good work out of it is mostly getting the right things in front of it at the right moment. Much of this book, the rules you will set next, the written plan you will produce in the next part, exists to keep the context that matters always in view.
There is also a ceiling on how much it holds at once, so a session that has run for hours goes vague and starts contradicting what you agreed earlier. Do not argue with it when that happens: open a fresh session and point it at the same files again. The build part turns this into a real skill.
1.2.4Trust it to type, not to decide
Let the agent do the typing, all of it, at full speed. That is what it is for, and second-guessing every line wastes the whole advantage.
But keep the decisions yours: what to build, whether the result is actually right, what is allowed to ship. Read what it produces instead of waving it through. The agent is the fastest pair of hands you will ever have, and you are still the one steering.
This prompt makes the agent show you what it actually is, on your own machine:
Do this now: paste the prompt and watch your agent read, run a command, and report back, so you have seen the loop before you start directing it for real.
