You already have Claude Code installed. The receptionist trainer and the Website Operator live in it.
Most people stop there. They open it for the business and close it two days later for themselves. Not because it cannot do the thing. Because nobody set it up for them.
The Knowledge Work Coach is $36, once, and its first job is the boring plumbing you would otherwise copy off a YouTube video wrong. Four words, because you will meet them everywhere from here on.
1. Hooks are tiny automations that fire at a moment
Think of the light that comes on when you open the fridge. That is a hook firing on "door opened".
Why do you need one? Because an AI model has no reliable clock. Every time you open Claude Code it wakes up with no idea what day it is, what timezone you are in, or what you were doing last night. Ask it to log something for today and it will confidently write a date from its training data.
The Coach wires two hooks.
A session-start hook runs the moment you open a session. It tells the model the date, the time and your timezone before you type a word.
A pre-tool-use hook fires right before the two logging tools the Coach uses, and hands them a fresh timestamp from your machine at that exact second. An event logged four hours into a session is stamped four hours later, not with the time you sat down.
Those two tools log tasks and goals to a dashboard. Purely analytical, purely for you, and none of your notes go anywhere. It is how you walk backwards across a week and see exactly what your Second Brain OS has been up to.
2. Git tracks changes line by line
You know the pattern. "final-v1", "final-final-v2", "ultra-final-v3". Developers solved that twenty years ago, and you have been living without the solution in Google Docs and Notion.
The Coach turns your knowledge folder into a git repository on day one. You never learn git. It just does it.
Here is why a folder of notes wants what developers use. Your notes are markdown, plain text with a little punctuation, so light that a year of daily notes weighs less than one photo. Plain text has lines. Claude Code does not swallow a 3,000-word document to change one sentence. It reads lines 40 to 60, edits line 52, done.
And then the biggest upgrade, one word: diff.
When Claude Code edits a line, you see the old line in red and the new one in green. Nothing else. There is no "I think it rewrote my intro". You look at four red lines and five green and you know. The model reads the same diff before it moves on.
Want to read a note like a human? Cmd + Shift + V in VS Code turns the raw markdown into the pretty version. Write in the raw, read in the pretty. No Obsidian, Notion or Evernote required.
One common confusion. Git is the technology. GitHub is a company that put it in the cloud. A GitHub backup is an option, not a requirement, and iCloud works too.
3. Memory is how you like things done
Not what you know. How you work.
"Put the date at the top." "My weekly plans live in the progress folder." "Never use em dashes." Every one of those is a preference, and without memory you state it again tomorrow.
Memory is a tiny folder of one-line files that Claude Code reads at the start of every session. That is the part of Claude Code that makes people give up after two days, and the Coach sets it up in the first hour.
4. QMD finds what you know
Memory is preferences. Knowledge is everything else: every reflection, every client call summary, every idea you wrote at 2am and forgot.
QMD stands for query markdown. It is a search engine that lives on your machine, points at your folders, and never touches the internet. Private, free, and fast.
The Coach installs it and registers each subfolder as a collection. From then on every note you save is indexed within seconds, and it searches three ways:
- By keyword. You know the word. "Stripe fee." Fastest.
- By meaning. You do not know the word. "That thing about payment costs eating margin." It finds the Stripe note anyway.
- By hypothesis. The model writes a fake paragraph of the answer and goes hunting for the note that looks most like it.
Combine keyword and meaning and your 2024 note surfaces next to your 2026 one, even if you never used the same words twice. And because it is local, the model reads the result by line number, exactly like git taught it.
Then it writes the way a second brain should be written
Hooks for automation, git for history, memory for preferences, QMD for search. Set up once, in the first hour.
After that, every piece of knowledge work becomes a markdown file with a title, tags and a date, in the collection it belongs to, on your disk. Searchable by meaning in under a second. And it logs what it did, so you can see exactly what happened from Monday to Sunday.
You get a knowledge-work employee you can monitor, not a chat window you have to babysit.
$36, once. All it needs is one hour of your uninterrupted attention.
A note on prices: every price in this series is the price inside its funnel, where each purchase unlocks the next. Bought on its own, outside the funnel, the same product can cost up to 30% more.
Next up: which laptop you actually need, and why it is not a Mac Mini.