One
A shopping list comes back wrong, because the file that lists what is in the kitchen is out of date.
The list leaves off broccoli. pantry.md still says a bag is in the freezer, and the bag was eaten a week ago.
Talk one of three · October 2026
What an agent is, and why git matters
This talk lands in the week of 12 October 2026.
AI in Systems Engineering · week 7 · Andrew Dunn
Scroll to run it Read on: the plates are printed at their finished state
Plate I
Most people patch the answer in front of them, and that correction dies when the session does.
One
A shopping list comes back wrong, because the file that lists what is in the kitchen is out of date.
The list leaves off broccoli. pantry.md still says a bag is in the freezer, and the bag was eaten a week ago.
Two
Two fixes are available, and on the day they are the same size.
Correct the answer: type the missing item into the reply. Correct the file: open pantry.md and delete one line. Write down which one you would do, before the weeks run forward.
Three
Six weeks later, only one of the two rooms has stopped seeing the mistake.
A correction typed into a reply lives as long as that conversation. A correction written into the file is read again next Sunday, and every Sunday after it.
Plate II
An agent is a language model given tools, running a loop: read, decide, act, observe, repeat.
One
A language model reads what is in front of it and writes text back.
Ask a chat window to delete the broccoli line and you get a sentence saying it has been deleted. Nothing on disk has changed.
Two
An agent is that same model placed in a loop and handed tools.
A tool is a named action the model may ask for, carried out by ordinary software: read a file, write a file, run a command. The model asks, the tool runs, the result comes back as the next thing the model reads. Read, decide, act, observe, and round again until it needs a person.
Three
One real tool carries today's example.
pandoc is a converter that turns one document format into another, and it is older than any of this. It turns a recipe web page into a plain file once: curl -sL URL | pandoc -f html -t markdown > recipes/SLUG.md. Every week after that, the agent reads the file instead of the page.
Plate III
Everything the agent knows fits one window, that window gets worse as it fills, and at the end of the session it goes to zero.
One
A context window is everything the model can see at once.
Your words, the files it read, the results of the commands it ran. The window has a fixed size, and inside one session the room it takes is never given back.
Two
It gets worse before it is full.
Roughly 40 percent of the window is where Dex Horthy puts the start of what he calls the dumb zone.
Past that mark the answers degrade: things read early get dropped, instructions get missed. The ink on the bar goes rough at the same place.
Three
Then the session ends, and all of it is gone.
The next session opens on an empty window. The only part that survives is what was written to a file, and the folder beside the bar is exactly as full as it was before.
Plate IV
The fix for a model that forgets is not a bigger window, it is a few plain files the agent reads before anything else.
One
A harness is the files and instructions around the model that decide what it knows and how it works.
This one is four markdown files, a folder of three recipe cards, and one command. Markdown is plain text with a few marks in it for headings and lists. There is no application here, and nothing in it is code.
Two
Five things, in the order the lesson builds them.
CLAUDE.md says who the agent is and what it must never do. The plan-the-week skill is the workflow, written as a document. pantry.md is what is on the shelves. preferences.md is memory. recipes/ holds the cards, one file per dinner.
Three
One line of the real preferences file reads: cauliflower in any form, the kids refuse.
Somebody wrote that down once, and no plan has proposed cauliflower since. All five are read at the start of the session, before the work starts.
Plate V
Once the fix is to edit a plain file, you need to see what changed and be able to take it back. git already does that.
One
The folder behind today's example is a git repository.
git is a tool that records a folder's history. A commit is one saved change to that folder, with a name, a date, and the person who made it. Every correction to a file arrives as one.
Two
preferences.md grows about a line a week.
Weeknight dinners should use at most one pan and one oven rack. If a recipe needs a marinade, save it for the weekend. Sumac, when a card offers sumac or za'atar. Each of these arrived on a Sunday, and the rail beside the card draws six of them.
Three
A correction can itself be wrong, so there has to be a way back.
The last entry on the rail takes one line back out of the file. None of this was built for AI. git is the ordinary tool underneath, and it turns out to be what a folder you correct by hand needed.
Plate VI
The loop only stays honest because it stops at fixed points for a person to decide.
One
The workflow runs in four phases: sort what we have, propose the week, build the shopping list, hold a retrospective.
It is one markdown file, and the agent works through it the way a person works through a checklist.
Two
Two of those phases stop and wait for a person.
The shopping list is not written until the week is approved. Nothing is written to memory until the diff, the list of what would change, is approved. The marker on the lane waits at each gate until you open it.
Three
What you do at a gate is the part a model cannot do for you.
Approve, and the plan stands as proposed. Tweak, and your correction is written into preferences.md, where next Sunday will read it. In this household it is about five minutes a week.
The marker is waiting at the first gate.
preferences.md; approve writes none.