not-ai-writer_skill

Overview

This skill is an expert guide for creating authentic, human-sounding content that avoids common AI-generated writing patterns. It helps you spot vocabulary, structural, tonal, and formatting markers that make text feel machine-crafted. Use it to review, edit, or create content that reads naturally and resists AI-detection pitfalls.

How this skill works

The skill runs a quick assessment checklist that scans vocabulary for overused AI words, flags formal transitions, checks sentence variety and paragraph patterns, audits voice and personality, and inspects formatting overuse. For each issue it identifies, it recommends targeted edits: simpler word choices, streamlined transitions, varied sentence rhythms, injected anecdotes or opinions, and more organic formatting. Actionable techniques and before/after examples illustrate each fix so you can apply them immediately.

When to use it

• Review copy suspected of triggering AI detection
• Edit AI-assisted drafts to sound genuinely human
• Write original content with a natural, conversational voice
• Polish marketing, docs, or emails to avoid robotic phrasing
• Troubleshoot content flagged by clients or platforms as AI-generated

Best practices

• Search and replace the top AI red-flag words with plain, conversational alternatives
• Limit formal transition phrases; let logical flow carry the text
• Deliberately vary sentence length and sentence-start patterns to increase burstiness
• Add specific examples, brief anecdotes, or confident opinions to inject personality
• Use formatting sparingly—mix narrative with selective bullets or subheadings

Example use cases

• Turn a corporate report into a readable summary with concrete examples and varied rhythm
• Edit blog drafts to remove AI vocabulary and add a distinct author voice
• Rework product documentation to be clear, specific, and human-focused
• Refine outreach emails to sound warm, direct, and less templated
• Improve social posts so they read like a real person, not a generated template