GOAL - W7 De-index untranslated/duplicate pages (SUBTRACTIVE) (odoo-expertos). Branch: fix/w7-deindex. Repo: C:\Users\eugen\.cursor\projects\Odoo Expertos\odoo-expertos. Work ONLY in this repo. Context: FULL-AUDIT-REPORT.md (Part B B0). Google flags ~103 pages as duplicates (PT/FR untranslated = duplicates of the Spanish original) and 587 crawled-not-indexed. This goal does NOT translate (humans do that later) - it removes the duplicate/thin pages from the index cleanly so they stop dragging the domain down. INPUT LISTS (already in repo): seo-audit-2026-05-27\gsc\duplicate-canonical.txt and seo-audit-2026-05-27\gsc\crawled-not-indexed.txt (URL lists from Search Console). TASK: For each URL in gsc\duplicate-canonical.txt that maps to a local /index.html: (1) add in its , and (2) remove its entry from the matching sitemap-*.xml. Do NOT delete the files (keep them for later translation). Do the same for clearly-untranslated non-ES pages you can confirm. Keep ES (root) pages indexable. CONSTRAINTS: only add the noindex meta + remove sitemap entries. Never noindex Spanish root pages or genuinely-translated pages. PROVE each turn: run powershell -File seo-audit-2026-05-27\REVIEW.ps1 -Only W7 and paste the full output. DONE WHEN: the output contains "W7: PASS" (sampled duplicate URLs still-indexable=0), AND you committed on branch fix/w7-deindex (show git log -1 --oneline). Stop after 25 turns if blocked and report the FAIL detail.