feat(content): consume use-cases.json from content repo (slice B) (#11)

Replaces the 280-line useCases.ts data array with a 30-line loader
that reads from InstaNode-dev/content/use-cases.json. UseCasesPage
unchanged.

Why JSON not markdown: each use case is a four-field record, not
prose — a structured format is the right tool. JSON also gives LLM
agents (Perplexity, ChatGPT search) an indexable, parseable surface
on GitHub in addition to the rendered HTML on instanode.dev. Two
surfaces for the same content.

Verified: 54 cards in dist/use-cases/index.html, all 9 categories
present, Ardent-style ephemeral-test-DB case still in slot 2.

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: mastermanas805