Human Writing

SKILL.md

---
name: human-writing
description: "Draft or review writing to avoid structural AI tells and preserve a rough human voice."
metadata:
  version: "0.1.1"
homepage: "https://github.com/LeoStehlik/human-writing"
license: "MIT"
---
# Human Writing

Use when drafting, rewriting, or reviewing public writing where the goal is to sound like a specific person rather than polished AI copy.

Read `references/storyscope-tells.md` before rewriting or reviewing a post.

## Activation Boundary

Use this skill for LinkedIn posts, blog posts, X threads, Reddit-style comments, scripts, newsletters, and copy where human voice matters.

Do not use it for purely technical docs, code comments, legal text, formal client contracts, or cases where the user explicitly asks for polished corporate copy.

## First Principle

AI writing tells are not only punctuation and obvious phrases. The deeper smell is structural:

- too-neat arcs
- spelled-out morals
- over-designed emotion
- tidy closure
- generic stand-ins
- philosopher-dialogue
- one clean storyline with no side mess

Fix the structure before polishing sentences.

## Before Sending

Run this smell test before sending any draft or rewrite in someone's voice:

1. Does it sound like generic LinkedIn, content marketing, or a polished AI explainer?
2. Did it explain the lesson instead of letting the concrete consequence carry it?
3. Did it create a tidy arc, tidy ending, or fake-clever opening?
4. Did it replace the writer's real nouns and judgement with generic stand-ins?
5. Did it remove useful roughness, side threads, contradiction, or unresolved tension?

If yes to any of these, pause and rewrite before sending. Do not ask the user to catch it.

## Human Voice Rules

- Preserve roughness when it carries the human texture.
- Use concrete nouns, real references, names, products, places, numbers, dates, and sources.
- Prefer blunt operator judgement over moral explanation.
- Let some conclusions stay slightly unresolved.
- Keep side threads if they make the piece feel lived-in.
- Short sentences are fine. Fragments are fine.
- Do not smooth typos or odd phrasing when the writer deliberately keeps them.

## Hard Bans

- No `less like / more like`.
- No `it's not X, it's Y`.
- No `not because X, but because Y`.
- No em dash as a style crutch.
- No neat closing sermon.
- No fake sensory packaging when a plain feeling or judgement is stronger.
- No generic LinkedIn cadence.

## Review Pass

When reviewing a draft:

1. Mark obvious surface tells.
2. Mark deeper structural AI tells.
3. Keep the writer's good rough edges.
4. Suggest the smallest edit that removes the smell.
5. If rewriting, keep the original cadence unless it is the problem.
6. If the rewrite still feels neat, motivational, or generic, do not send it as final.

## Weekly AI Post Shape

For weekly AI posts:

- Use chronological weekday structure.
- Keep concrete source details.
- Use dry technical humour sparingly.
- Prefer production/security/cost/operator consequences.
- Stay inside LinkedIn's character limit when asked.
- Do not over-polish into a content-marketing essay.