War-Story Notes
How to write the notes your reply engine can actually quote — short, first-person, verifiable facts that turn your real work into grounded replies.
Your reply engine has one unbreakable rule: every drafted reply must stand on a fact quoted verbatim from your own notes, verified character for character. That rule is what keeps your replies out of the AI-slop pile — and it has a consequence most people discover the hard way. The engine can only be as good as the sentences it's allowed to quote.
A war-story note is a note written specifically to be quoted: short, first-person, verifiable facts about what you've actually done. One per line. It is the highest-octane fuel the reply engine takes.
Why your existing notes probably don't work
Most knowledge bases are full of specs, plans, and documentation — paragraphs about things. When the anchor extractor reads a spec, the best quote it can find is jargon: mode names, configuration details, feature lists. Those either fail the quality gates or produce replies that sound like a manual.
Watch the difference:
Spec-shaped (fails): "The system supports opt-in whole-vault indexing mode with safety checks for out-of-scope writes."
War-story-shaped (anchors a great reply): "I spent twenty years not finishing blog posts. Last month I published nine."
Both are true for their authors. Only one sounds like a person, survives being quoted without context, and gives a stranger a reason to believe you.
The format
Create one note per theme (your craft, your business, your system — whatever your Hunts are about) and fill it with single-sentence facts. Each line must be:
- One sentence, self-contained. It will be quoted verbatim with no surrounding context. If it needs the line above it to make sense, it can't be used.
- First person. "I rebuilt the store's checkout in a weekend," not "checkout was rebuilt."
- Hard to believe and easy to verify. The best receipts carry a number, a timespan, a before/after, or a limit you hit honestly. "I've onboarded 40 clients" beats "I have extensive client experience" every time.
- In your own voice. These lines land inside your replies word for word. If it reads like marketing copy in the note, it reads like marketing copy on X.
- True without asterisks. The engine's whole doctrine is no unsupported claims. Never round up, never borrow, never predict.
What qualifies
- A number that got bigger: clients, posts shipped, years in the trade, revenue you'll say out loud.
- A failure you fixed and what it cost you first.
- A thing you refused to do, and why — restraint is a credential.
- A tool, habit, or process you built for yourself that solved a real problem.
- A limit you hit honestly: "this didn't work until…" is more credible than any success story.
What never qualifies: aspirations, opinions without a story behind them, anything you'd have to soften if a stranger asked "prove it."
A worked example
A fractional CFO's war-story note might read:
- I've closed the books for 30 companies and 11 of them had the same broken revenue-recognition mistake.
- I turned down a client last year because their numbers couldn't support the raise they wanted, and told them so in the first meeting.
- My month-end close checklist is 41 steps because every one of them is a mistake I made once.
Every line is quotable alone, verifiable, and answers a conversation happening on X right now.
How it compounds
Add the note to a Brain and bind that brain to your Hunts. From then on:
- The anchor extractor quotes your receipts verbatim to ground daily queue drafts.
- Every time you ship, fix, or learn something, append one line. A war-story note that grows weekly means a reply engine that gets more specific weekly.
- Thin days in your queue are usually a receipts problem, not a system problem. When your build note says "held back (no anchor fact)," the fastest fix is five new lines in this note.
Write the first ten lines in ten minutes: open the note, finish the sentence "Most people don't believe me when I tell them…" ten times, and keep every answer you can prove.
Related
- Fact-Anchored Replies — how the verbatim-quote guard uses these lines
- Reply Cadence Engine — the daily queue these notes feed
- Ranked Reply Targeting — the Hunts that find the conversations your receipts answer