# [Name], voice file

**Built from:** last [N] LinkedIn posts, [first date] to [last date], pulled [date] from your own activity page. [N] regular posts, [N] lead magnet posts.
**Refresh:** once a month, or after any month you post a lot.

## How Claude uses this file

- Read all of it before drafting anything.
- Every rule with a count was measured on your posts. Don't break one without a reason you could say out loud.
- After drafting, run `check_draft.py` against this file and fix every line it flags before showing the draft.
- When you reject a line Claude wrote, it adds a row to the ban list at the bottom.

## Who's writing

Three lines, taken from the posts, not from a bio. What you do, who you write for, what you keep coming back to.

## Rules you can check

| Rule | What the posts show |
|---|---|
| [e.g. Em dashes] | [count] in [N] posts |
| [e.g. Opening lines that are questions] | [count] of [N] |
| [e.g. P.S.] | [count] of [N] lead magnet posts, [count] of [N] regular posts |
| [e.g. Emoji] | none in [count] of [N] posts |
| [e.g. Lists] | [marker] in regular posts, [marker] in lead magnet posts |
| [e.g. Length] | regular posts median [N] characters, lead magnet posts [N] |

Only rules a script or a reader can verify go in this table. "Casual tone" is not a rule. "lol in 5 regular posts, never in a lead magnet post" is.

## How your posts open

| Type | Posts | Median reactions | Median comments | Your words |
|---|---|---|---|---|
| [type] | [N] | [N] | [N] | "[first line, verbatim]" |

Sorted by how often you use each one. The results columns say which ones to use more.

## Lines only you would write

10 to 15 sentences copied word for word from your posts, with the date. Pick lines a stranger couldn't have written: specific numbers, your opinions, your jokes, the way you swear or don't.

## How your posts end

The last lines you actually use, by type, with counts and examples.

## Regular posts vs lead magnet posts

What changes between the two: length, lists, emoji, P.S., register.

## Register

The casual words you use and where (regular posts only, or everywhere). Words in caps for emphasis and how often. Typos you leave in, if any.

## What's working

Your best and worst regular posts from the pull, with reactions, comments and impressions. One line on what the best ones share.

## Ban list

Every line you've rejected, and why. Claude adds a row each time you say a line doesn't sound like you. `check_draft.py` flags anything in the first column; wrap a pattern in /slashes/ to match it as a regex.

| Pattern | Line you rejected | Why it's not you | Date |
|---|---|---|---|
| "[phrase]" | "[the full line]" | [one line] | [date] |

## Machine-readable rules

Paste the block `voice_stats.py` printed. `check_draft.py` reads it.

```rules
posts: 0
median_chars_regular: 0
median_hook_chars: 0
em_dash_posts: 0
fake_bold_posts: 0
question_hooks: 0
ps_on_regular: 0
no_emoji_posts: 0
```