The agent loop
By the end of today you can write the agent loop from memory as ordinary code, name every place it can fail, and explain why an 'agent' is a while-loop around a model that can call your functions rather than a new category of software.
YesterdayOn Day 77 you let a model choose one of your functions. On Day 78 you validated its arguments and handled tool errors.
TomorrowTomorrow you'll ask the harder question: when a deterministic workflow beats an agent, and why that's usually the right answer.
Why this matters
Agent frameworks hide about forty lines of control flow. Engineers who have never seen those forty lines cannot debug an agent that loops forever, calls the wrong tool, or silently burns $80 in tokens.
- The agent loop
- Agent termination
- Tool result feedback
- Loop budget
- Agent failure modes
Learn it
80 minCopy this into Claude or ChatGPT. It quizzes you before it explains anything, which is deliberate. The resources under it are how you check what it told you.
Today's Master Prompt
Free · sign inA prompt written for this day alone: your level, the exact scope, what to leave out, and an instruction to quiz you before it explains anything. Paste it into Claude or ChatGPT and it teaches you today's material.
Check it against something that is not a model
An assistant can be fluent and wrong, and on a topic you met today you will not catch it. These cover the same ground and were made by people who do this for a living, so they are what you hold the explanation up against. They are other people's work and we only link to them, so judge them for yourself.
4 hand-picked resources
Free · sign inVideos, official docs and articles covering the same ground, each opened and annotated by hand. They are what you check the assistant against on a day you cannot yet catch it being wrong.
Build it
50 minWrite the agent loop yourself in about 40 lines, with no framework. Give it two tools of your choosing. Add a hard iteration cap and a token budget, and log every turn to stdout so you can watch it think. Then deliberately break it: make one tool always raise an exception, and watch what the model does across turns.
Recall it
20 minAnswer out loud, reveal, then mark honestly whether you had it. That score is the only thing on this page you do not get to choose.
5 recall questions
Free · sign inQuestions you answer from memory, then grade yourself against the real answer. The score is carried into the mastery rating below it, so an honest miss cannot quietly become a tick.
Rate it
Completion and mastery are tracked separately. Be honest, because an inflated rating only means the concept resurfaces sooner.
Mastery tracking
Free · sign inRate yourself against five named criteria per concept. Completion and mastery are tracked separately, and anything you rate shakily comes back automatically on a spaced schedule.
Recap
- 01An agent is a loop around a model that can call your functions
- 02The loop exists so the model can see results before deciding again
- 03Never trust the model alone to terminate; cap iterations, time and tokens
- 04Cost grows quadratically because the full history is re-sent every turn
Your progress
Free · sign inMark days complete, pick up where you left off across devices, and watch completion and mastery diverge. Free, and the account exists only so ninety days of work cannot vanish with a cleared browser.
Extra time
OptionalOptional. Nothing in a later day depends on anything here.
If you have another hour
Now read the source of a small agent framework and map its abstractions back onto the forty lines you wrote. Find its loop. Find its termination guards. Find where it builds the message history. The goal is to see that the framework did not invent anything; it only named things you already wrote.