Summary

Adds avoid-ai-writing to the Communication & Writing section.

What problem it solves

AI-generated text contains identifiable patterns — em dash overuse, copula avoidance, significance inflation, synonym cycling, and more — that undermine credibility. This skill systematically identifies and fixes them. It audits content against 21 categories of AI writing patterns and applies a 43-entry replacement table to produce output that reads like a person wrote it.

Who uses this

Anyone using Claude to draft content — blog posts, documentation, marketing copy, internal communications — who wants the output to read like a person wrote it.

Attribution

Pattern research based on Wikipedia's "Signs of AI writing" documentation and the blader/humanizer Claude Code skill.

Example

Paste any text and get back four sections:

1. Issues found — quoted excerpts with the specific AI pattern identified
2. Rewritten version — the full text rewritten with all patterns removed
3. Change summary — a table of every change made and why
4. Second-pass audit — a re-read of the rewrite that catches any surviving tells

Test plan

• Entry added in alphabetical order within Communication & Writing section
• Markdown link format matches existing entries
• Description follows the style of other entries in the list