Can I use this for academic literature reviews?

Yes, version 3.3 includes specific patterns and interview rounds optimized for academic writing, critiques, and research summaries.

How does Guided Interview Mode help avoid AI detection?

By asking you questions first, the skill ensures the core phrasing and 'soul' of the text originate from you, creating the high-perplexity sequences that signal human writing to detectors.

What is the 'Perplexity Ceiling' mentioned in the skill?

It is the discovery that when an AI edits AI-generated text, the statistical fingerprint remains detectable. This skill solves this by prioritizing human-first drafting.

Does this skill simply replace common AI words?

No, it goes beyond word swapping. It uses a Guided Interview Mode to extract human thoughts, as statistical classifiers detect token probability distributions that simple editing cannot hide.

AI Writing Humanizer

Name: AI Writing Humanizer
Author: Touricks

byTouricks

•

Gestión de Contenido

Eliminates AI-generated writing patterns using a human-first guided interview workflow to ensure text passes statistical classifiers.

The Humanizer skill is a sophisticated editing utility designed to strip away the predictable statistical fingerprints of LLM-generated text. Moving beyond simple word replacement, it addresses the 'Perplexity Ceiling' by implementing a unique Guided Interview Mode. This mode uses interactive Q&A rounds to extract a user's genuine thoughts, vocabulary, and emotional reactions before any writing begins, ensuring the final output is built upon a foundation of human-origin token sequences. It is particularly effective for academic reflections, paper critiques, and professional documentation where passing AI detectors like GPTZero is a priority.

Características Principales

01Detection and removal of Wikipedia-defined AI writing patterns

02Academic-specific pattern matching for literature reviews and critiques

03Refined pattern checking for 'second-generation' AI tells and formulaic structures

04Statistical perplexity injection to lower AI detection probability

05Guided Interview Mode to extract authentic human thoughts and vocabulary

062 GitHub stars

Casos de Uso

01Humanizing AI-assisted professional reports to sound more natural and less formulaic

02Creating academic critiques and reflections that reflect a genuine personal perspective

03Drafting content that bypasses statistical classifiers like GPTZero and Originality.ai

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add touricks/fanshi_personal_skills humanizer

For use in Claude.ai and ChatGPT

Características Principales

01Detection and removal of Wikipedia-defined AI writing patterns

02Academic-specific pattern matching for literature reviews and critiques

03Refined pattern checking for 'second-generation' AI tells and formulaic structures

04Statistical perplexity injection to lower AI detection probability

05Guided Interview Mode to extract authentic human thoughts and vocabulary

062 GitHub stars

Casos de Uso

01Humanizing AI-assisted professional reports to sound more natural and less formulaic

02Creating academic critiques and reflections that reflect a genuine personal perspective

03Drafting content that bypasses statistical classifiers like GPTZero and Originality.ai