This generates content optimized for AI citations in ChatGPT, Claude, and Gemini rather than traditional search rankings. You'd use it when creating FAQ schemas, analyzing pages for LLM extraction, or building evidence panels that AI models actually quote. The core insight is the 18-token extraction rule: LLMs pull self-contained sentences of roughly 15-20 words, so your claims need to stand alone without context. It pushes single-topic pages over comprehensive guides, requires citations on everything, and treats authority level as strategy (aggressive optimization for new sites, light touch for established ones). Includes templates for FAQPage schema, evidence panels, and a testing protocol across major AI models.
npx skills add https://github.com/bencium/bencium-claude-code-design-skill --skill bencium-aeo