[voltcrash]

Search engines and AI search tools received an empty JavaScript app container and a sitemap of destination query URLs sharing the same initial metadata. Exora now publishes useful, source-linked HTML that explains the application and gives people a direct route into its 3D destinations.

• Publish 28 indexable pages: the homepage, About, discovery indexes, five practical guides, 16 NASA-backed planet profiles and three existing tour itineraries. Profiles preserve unknown values, property provenance and the snapshot date.
• Generate canonical URLs, social previews, WebSite/WebApplication/page/breadcrumb JSON-LD, a sitemap of readable pages and an optional llms.txt directory. Keep search and AI retrieval crawlers allowed under the existing open policy; exclude API responses from indexing.
• Link discovery resources from Explore, isolate their offline cache entries from the explorer shell, provide a static missing-page response, and validate emitted metadata and internal links on every build.

Proof: vp check; vp test (595 tests); recursive workspace tests and builds; asset provenance and production dependency audit; 74 desktop/mobile browser tests (38 existing skips); no-JavaScript desktop/mobile page checks and the About → explorer → Explore → Guides browser journey.

Lighthouse passed all enforced desktop/mobile assertions on the homepage, About, exoplanet guide and TRAPPIST-1 e profile. All eight reports scored 100 for SEO; the three sampled discovery pages also scored 100 for performance, accessibility and best practices on both devices. The interactive homepage still records performance warnings (67 desktop / 63 mobile, with main-thread work and unused JavaScript); SEO is now enforced at 100 without weakening other checks.

Decisions: use the existing catalog and tour definitions without adding dependencies; deliberately exclude duplicate interactive query URLs from the sitemap; preserve current destination sharing and canonical behavior. Search Console/Bing submissions and hosting firewall verification need verified external accounts and are documented in docs/search-discovery.md. Rankings and AI recommendations cannot be guaranteed; llms.txt is an optional content directory, not a ranking claim.

Review notes: simplified pathname routing and trimmed duplicated unit assertions. Kept the API header test consistent with existing config tests, retained self-canonical slash URLs without a global redirect affecting APIs, and intentionally replaced query-only sitemap entries with readable resources. One review round; no blockers or should-fix findings. Reviewer usage was not reported. Final diff: 24 files, 1,111 additions / 87 deletions; review removed five duplicate assertion lines and unnecessary query parsing.

Reviewed with review-loop: 1 round with Claude Opus 5.5. Reviewer verdict: approved.

Created with GPT-6.1 Sol in T3 Code through the Codex harness.