search-optimization-architectlisted

# Skill: Search Optimization Architect — Retrieval, Extraction, Citation > **Apply Guidelines Skill** — load the `guidelines-meta` skill before proceeding. > **Writing floor:** every content artifact this skill emits is a document. Apply `module-writing-floor` to it. This skill owns *what makes it retrievable*; that module owns *how it reads*. > **Design floor:** any page this skill restructures is still a page. Apply `module-craft-floor` before shipping visible changes — a retrieval-optimized page that reads as machine feed has failed the human half of its job. > **Profile section owned:** §Search Visibility (Guidelines §5). On first use, if it is missing or `TODO`, **read the repo and the live site first** — the rendering mode is in the framework config and in `view-source`, the robots policy is in `robots.txt`, the schema is in the emitted `<script type="application/ld+json">`, the sitemap is at its declared path, the analytics property is in the tag config. Then fill it per **Guidelines §5.1–§5.4**. **The two failures this exists to prevent:** 1. **The page is written for a reader who never arrives.** A retrieval pipeline does not read pages. It fetches whatever HTML the first response contains, splits it into chunks, embeds them, retrieves a handful, and synthesizes. Everything that depends on the rest of the page — a pronoun resolved three paragraphs up, a definition in the hero, a number in a chart, a claim rendered by JavaScript — is gone by the time the model sees