Deslopify Skill

A specialized skill for Gemini CLI that makes text more genuine, natural, and free of "AI slop" by identifying and removing common LLM writing tropes.

Goal

The goal of this skill is to transform AI-generated or overly formal text into clear, direct, human-like prose. It helps you move away from the repetitive, patronizing, and grandiose patterns that often signal a piece was written by a machine.

How it Works

The skill is grounded in the comprehensive AI Writing Tropes list. It scans your text for anti-patterns such as:

• Magic Adverbs: Overuse of words like "quietly," "deeply," or "arguably."
• The "Delve" Family: Overused vocabulary like "leverage," "streamline," or "harness."
• Pompous Constructions: Patterns like "It's not X -- it's Y" or "The result? Devastating."
• Structural Tics: Bold-first bullets, em-dash addiction, and tricolon abuse.
• Tone Issues: Patronizing analogies ("Think of it as...") and grandiose stakes inflation.

Usage

Once installed and reloaded, you can use the skill by asking Gemini CLI to "deslopify" or "naturalize" a piece of text or a URL.

Example Commands:

• "Deslopify this article: [URL]"
• "Naturalize this draft: [Text content]"
• "Remove the AI slop from my README file."

Installation

If you have the .skill file, you can install it using:

gemini skills install deslopify.skill --scope workspace

Then, reload your skills in your interactive session:

/skills reload

Reference

This skill's stylistic rules are strictly based on the observations and research found at tropes.fyi.