
from ai-tells-validator17
Detects and removes common 'AI-sounding' linguistic patterns and banned words using a live Wikipedia-grounded catalog.
This skill provides a deterministic system for detecting 'AI tells'—linguistic markers, banned words, and structural patterns that typically signal LLM-generated content. It allows agents to validate prose for human-like quality and automatically rewrite drafts until they pass the validation check.
Use this skill when preparing high-stakes external communications where authenticity is critical, such as:
validate.mjs for detecting tells and rewrite.mjs for iterative cleaning via LLM.Designed for agents capable of running Node.js scripts, specifically optimized for Claude Code.
This skill has not been reviewed by our automated audit pipeline yet.