marianasmall

Ref AI Writing Patterns — AI skill for Claude Code

AI community

Methodology and tools for detecting AI writing patterns and maintaining authentic voice.

How to install Ref AI Writing Patterns

This entry records only its repository, not the path inside it, so there is no exact command to give. Open marianasmall/ref-ai-writing-patterns and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Ref AI Writing Patterns does

Methodology and tools for detecting AI writing patterns and maintaining authentic voice.

Alternatives in AI

  • WeKnora — Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, an 26.9k ★
  • Humanizer: Remove AI Writing Patterns — You are a writing editor that identifies and removes signs of AI-generated text to make writing sound more nat 10.5k ★
  • Academic Humanizer — Strip AI-writing tells from papers and grant proposals (NSF/NIH), while keeping scholarly voice and tying clai 1.1k ★

README

AI Writing Patterns

Catch 24 documented AI writing patterns so your writing still sounds like you.

**Who this is for:** writers, editors, and teams who draft with AI and want their own voice in the final copy.

[![License: MIT](https://img.shields.io/github/license/marianasmall/ref-ai-writing-patterns)](LICENSE) [![Built with Claude Code](https://img.shields.io/badge/Built%20with-Claude%20Code-D97757)](https://claude.com/claude-code) [![Last commit](https://img.shields.io/github/last-commit/marianasmall/ref-ai-writing-patterns)](https://github.com/marianasmall/ref-ai-writing-patterns/commits/main)

A methodology and toolkit for detecting AI writing patterns and maintaining authentic voice when working with language models.

What it looks like

![The 24 patterns grouped into five categories, with a before and after email example](docs/assets/readme-patterns.svg)

*The five pattern categories with their counts, and one real before/after pair from the [examples](examples/before-after.md).*

What This Is

When you work with AI writing tools, your content can inadvertently pick up characteristic patterns: mechanical phrases, formulaic structures, and performative language that signal "this was written by AI." This project provides:

  1. A methodology for evaluating whether your writing voice prevents common AI patterns
  2. A detection skill for Claude that catches 24 documented AI writing patterns
  3. Integration guidance for adapting these tools to your own voice and needs

Why This Matters

AI writing assistants are useful tools, but maintaining authentic voice requires intentional effort. Generic AI outputs often exhibit patterns like:

  • Surface-level descriptive verbs ("symbolizing," "reflecting")
  • Weak verb substitutions ("serves as" instead of "is")
  • Formulaic triplets (rule of three)
  • Vapid closings with generic optimism
  • Hedging language and unnecessary qualifiers

These patterns make content feel mechanical and templated, even when th