[tiXor-code]

TLDR: 'termoenergetica avarii' gets 4,592 impressions/28d at position 9.7, but Google splits it across /sector/6, /sector/1 and the contact guide because no page owns the city-wide intent. This PR adds /avarii, a city-wide roll-up of the ongoing Termoenergetica announcements, and links it from the nav, homepage, sector pages, sitemap and llms.txt.

Why

GSC query-page data (last 30d) for 'termoenergetica avarii': /sector/6 (2,435 imp, pos 9.6), /sector/1 (1,161 imp, pos 9.9), contact guide (538 imp, pos 10.2, 0 clicks). Sector pages answer the per-sector version of this query; nothing answers the city-wide version. Same pattern the site already uses for map queries (/harta owns them, the homepage links there).

What changed

• New app/avarii/page.tsx: answer-first dated status sentence (count of ongoing interruptions, how many are avarii), the operator's announced restore terms (marked as operator estimates, never our forecast), the grouped announcements table with a Sector column (same rendering idiom as /sector/[id]), a per-sector summary table (ongoing / avarii / PTs hit in the last 30 days), a contact section (TELVERDE, 031.9442, cmteb.ro, TERMOALERT, all repeating facts already sourced in the contact guide), and a 5-question FAQ with FAQPage JSON-LD via lib/seo.ts faqJsonLd.
• City data comes from the existing getSectorLive per sector, rolled up at build time; no new data-layer code, pure SSG preserved (dynamic = 'error').
• Internal links to the new page: SiteNav ('Avarii azi'), homepage 'Avarii azi' bullet, sector page footer ('Toate sectoarele: avariile de azi'), sitemap.ts, llms.txt.

Verification

• npm test: 137/137 passing.
• npm run build: compiled clean, 6,335 static pages including /avarii; the one Turbopack NFT warning pre-exists on main (verified by building main).
• Rendered /avarii HTML inspected: status sentence, both tables, FAQPage JSON-LD with 5 questions, breadcrumb JSON-LD all present; copy uses proper diacritics and the dated-snapshot framing (never implies real time).

Order: SEO Sentinel content/brief id 121.

Generated with Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
termo-site	Ready	Preview	Nov 2, 2026 8:50am UTC

[claude]

Reviewed the full diff plus the lib it depends on (lib/sector-live.ts, lib/seo.ts, lib/format.ts, lib/data.ts).

• Correctness: getCityLive rolls up getSectorLive per sector correctly; group sort (avarie-first, newest-first) matches /sector/[id]; statusSentence/restoreSentence/metaStatus branch consistently for the n=0/1/mixed/all-avarie/all-programat cases. The /ghid/care-sunt-informatiile-de-contact-pentru-avarii-la-termoenergetica link target exists in content/ghid/.
• Security: FAQPage JSON-LD goes through JsonLd in lib/seo.ts (the required < escaping is intact), and the table renders scraped PT/street names as plain JSX text (React-escaped), not via dangerouslySetInnerHTML — consistent with the CLAUDE.md rule on untrusted CMTEB input.
• SSG constraint respected: dynamic = 'error', no client components, data read at build time only.
• Minor non-blocking nit: generateMetadata() and the page body each independently call getCityLive(lcy), so getSectorLive scans all 6 sectors twice per build of this one page — harmless for a single static route, not worth blocking on.

Nav/homepage/sector-footer/sitemap/llms.txt link additions are small and consistent with existing idioms.