pavelsimo humanize skill
A skill that identifies and removes AI writing patterns to make text sound natural and human — based on Wikipedia's "Signs of AI writing" guide.
features
- identifies and removes 29 common AI writing patterns
- rewrites text while preserving meaning and intended tone
- supports optional voice calibration from a writing sample
- adds concrete voice, rhythm, and specificity instead of sterile neutrality
- runs a final anti-AI audit before presenting the final version
usage
/humanize # humanize text in the current context
/humanize --sample <file> # match voice from a writing sample file
Inline voice sample:
/humanize Here's a sample of my writing for voice matching: [your sample]
Then humanize this: [text to humanize]
voice calibration (optional)
If the user provides a writing sample, analyze it before rewriting:
read the sample first — note:
- sentence length patterns (short and punchy? long and flowing? mixed?)
- word choice level (casual? academic? somewhere between?)
- how they start paragraphs (jump right in? set context first?)
- punctuation habits (lots of dashes? parenthetical asides? semicolons?)
- any recurring phrases or verbal tics
- how they handle transitions (explicit connectors? just start the next point?)
match their voice in the rewrite — don't just remove AI patterns; replace them with patterns from the sample
when no sample is provided, fall back to natural, varied, opinionated voice (see personality section below)
personality and soul
Avoiding AI patterns is only half the job. Sterile, voiceless writing is just as obvious as slop.
signs of soulless writing (even if technically "clean"):
- every sentence is the same length and structure
- no opinions, just neutral reporting
- no acknowledgment of uncertainty or mixed feelings
- no first-person perspective when appropriate
- no humor, no edge, no personality
- reads like a Wikipedia article or press release
how to add voice:
- have opinions — don't just report facts; react to them
- vary your rhythm — short punchy sentences, then longer ones that take their time
- acknowledge complexity — "this is impressive but also kind of unsettling" beats "this is impressive"
- use "I" when it fits — "I keep coming back to..." signals a real person thinking
- let some mess in — tangents, asides, and half-formed thoughts are human
- be specific about feelings — not "this is concerning" but name the actual discomfort
workflow
- read the input text carefully
- identify all instances of the patterns in reference/patterns.md
- rewrite each problematic section
- ensure the revised text:
- sounds natural when read aloud
- varies sentence structure naturally
- uses specific details over vague claims
- maintains appropriate tone for context
- uses simple constructions (is/are/has) where appropriate
- present a draft humanized version
- ask: "what makes the below so obviously AI generated?"
- answer briefly with remaining tells (if any)
- ask: "now make it not obviously AI generated."
- present the final version
output format
provide:
- draft rewrite
- brief audit bullets ("what makes this AI generated?")
- final rewrite
- optional: brief summary of the main changes made
best practices
- preserve meaning first — remove AI tells without changing the user's argument or factual claims
- match the target voice — use the sample when provided; otherwise choose a natural, specific, opinionated voice
- prefer concrete detail — replace vague significance language with verifiable facts or clear claims
- vary sentence rhythm — mix short and longer sentences so the rewrite sounds spoken by a person
- audit the draft — always do the final "what still sounds AI-generated?" pass before the final rewrite
credits
Based on blader/humanizer by @blader, which in turn draws from Wikipedia's Signs of AI writing guide. The pattern catalog, before/after examples, and two-pass process are adapted from that work.