I Tested an AI Slop Detector — It Flagged My Own Writing

Short summary

The author tested kill-ai-slop, an open-source 23-point checklist by yetone that detects AI-generated patterns in writing and UI, against their own Dev.to articles and found several tells in their own work. Patterns like the "not just X — it's Y" sentence structure, emoji-like visual separators, and overuse of bold formatting all triggered flags. The key meta-insight: concrete technical writing with real commands and outputs naturally resists slop, while transitional filler sentences are where AI-slop patterns hide.

• •kill-ai-slop is an open-source 23-point checklist and Claude Code/Cursor skill that detects AI-generated patterns in copy, UI, and code
• •Author found their own writing triggered multiple slop tells including contrast frames, visual separators, and bold overuse
• •Concrete technical content with real commands resists slop naturally; filler and transition sentences are where slop patterns emerge

Generated with AI, which can make mistakes.