A comprehensive reference for spotting AI-generated text, built from academic research and real-world detection patterns. Covers vocabulary fingerprints like "delve" and "tapestry," structural tells like uniform sentence length and tricolon abuse, plus model-specific artifacts (ChatGPT's turn0search markers, Grok's XML tags). The methodology is layered: scan for technical artifacts first, then analyze vocabulary clusters, structure, and citations. Includes false positive prevention guidance since non-native speakers get flagged at 61% rates. Useful when you need to verify authorship or understand how detection actually works beyond vague "it sounds robotic" intuition. Requires 200+ words and multiple signals for reliable results.
npx skills add https://github.com/mike-coulbourn/claude-vibes --skill ai-writing-detection