humanizer

Rewrite AI-sounding prose to remove detectable AI writing patterns while preserving meaning.

Updated Sep 8, 2026
One-click install
npx skills add https://github.com/salomepoulain/makery-stations --skill humanizer-salomepoulain
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: humanizer
Source: https://github.com/salomepoulain/makery-stations/tree/main/stations/claude/workbench/pantry/skills/humanizer
Command: npx skills add https://github.com/salomepoulain/makery-stations --skill humanizer-salomepoulain

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI-generated text carries recognizable patterns—staged openers, not-X-but-Y contrasts, forced triads, inflated claims, and chatbot residue—that make prose sound generic and machine-written. This Skill rewrites such text so it reads like a human wrote it, without changing what the text says. ## Core Features & Use Cases - Pattern detection and removal: Identifies 25 documented AI writing tells drawn from Wikipedia's "Signs of AI writing," ranked by strength, from chatbot residue to decorative bolding. - Voice matching: When given a writing sample, matches its sentence length, word choice, punctuation, and transitions so the rewrite sounds like the original writer. - Fact preservation: Keeps every supported claim, name, number, date, and citation while removing stylistic tells; never invents facts or sources. - Use Case: Paste a blog draft that reads like chatbot output—full of "Let's dive in," em dashes, and one-line closers—and get back a rewrite that states each point directly in a natural voice, plus a list of remaining patterns. ## Quick Start Rewrite this pasted article to remove AI writing patterns while keeping all facts and matching my usual tone.

Frequently Asked Questions about humanizer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I make AI-generated text sound more human?

Identify and remove structural AI tells rather than swapping individual words: staged openers, not-X-but-Y contrasts, one-line closers, forced triads, and inflated significance claims. State each point directly, vary sentence length, and keep only sentences that add new information.

What are the most common signs of AI writing?

The strongest tells are chatbot residue like "I hope this helps," not-X-but-Y contrasts, dramatic one-line closers, and staged run-ups like "Let's dive in." Weaker signals include overused em dashes, forced triads, bold labels on every list item, and stock words like "delve" and "landscape."

Does rewriting AI text change the facts or meaning?

No. The rewrite keeps every supported claim, name, number, date, quote, and citation from the source. It never adds facts not present in the source or supplied by the user, and treats any unsupported addition or lost claim as an error.

Can the rewrite match my personal writing style?

Yes. Provide a writing sample and the rewrite will match its sentence length, word choice, punctuation habits, openings, and transitions. The sample overrides default rules, so if you normally use dashes, they are kept at your usual rate.

When should AI writing patterns be left alone?

Leave watched phrases inside quotations, titles, proper names, and passages discussing the phrase itself. Text written before November 2022 is not AI-written, and weak tells like a single dash or curly quotes only justify edits when several tells appear together.