feat(library): add FAQPage JSON-LD to the eight keyword-research spokes

[jeremypenguin-max]

What

Adds FAQPage JSON-LD to the eight keyword-research library spokes — 27 questions total. No other changes.

Spoke	Q&A pairs
seed-from-conversation	3
long-tail-question-mining	3
search-intent-mapping	3
cluster-topical-hubs	3
gsc-programmatic-discovery	4
opportunity-sizing-forecasting	4
intent-beyond-google	4
positioning-to-demand	3
Total	27

Each file gets the same three things, copying the shape already used on the pillar (library/keyword-research/index.tsx): a faqs array, a faqLd object, and one dangerouslySetInnerHTML script. The component body is wrapped in a fragment so the script can sit beside <LibrarySpokePage>.

Why

The FAQ content already exists and is already visible on every one of these pages — a ## … FAQ H2 with ### questions and their answers, authored in the MDX. Only the markup was missing. The pillar page has emitted FAQPage since it shipped; the spokes never did.

How it was verified

The rule that matters here is that FAQ markup must match what is actually rendered — markup Google cannot see on the page is a violation, so a mismatched block would be worse than no block at all. So none of these strings were retyped:

1. Extracted programmatically from the MDX. Question from the ### heading, answer from the paragraph beneath it. Inline markdown was resolved to the text a reader actually sees — inline-code backticks stripped (`noindex` renders as noindex), and link syntax reduced to its anchor text.
2. Checked against the prerendered HTML, not the source. After vite build, every application/ld+json block was pulled out of each built page, JSON.parsed, and then every question and answer string was searched for verbatim inside the tag-stripped visible text of that same page.

Result: 27/27 pairs matched verbatim, 8/8 pages. Every page carries exactly two ld+json blocks — the new FAQPage and the BreadcrumbList that LibrarySpokePage already emitted — so there is no duplicate breadcrumb.

Checks:

• npm --prefix web run types:check — clean
• npx vite build — all eight spoke URLs prerender, no errors
• npx prettier --check on the eight touched files — clean
• Every JSON-LD block JSON.parses out of the built HTML

Scope notes

• No prose changes, no new pages, no new dependencies.
• No breadcrumb work: LibrarySpokePage already emits BreadcrumbList for these routes.
• Does not depend on, and is not stacked on, any other open PR.
• The same gap still exists on the competitive-analysis and site-audit spokes. Those live in unmerged PRs, so they are deliberately left out of this one; they can be covered once those merge.

🤖 Generated with Claude Code