Reads existing brand DNA, keywords.csv, and prompts.csv, then produces a plan.csv — a strictly-schema'd content architecture telling the next pipeline step which pages to create, for which keyword/prompt clusters, and in what build order.