Every index accumulates junk over time: parameter URLs from sorting, internal search results, print versions, pagination duplicates — pages that were never meant to show up in search. Individually harmless, together they dilute how a search engine reads site quality and pull crawl budget away from pages that actually deserve to rank.
How duplicate pages and index junk are found
Given a domain, the skill checks the overall size of its index as a baseline, then searches for common junk patterns — parameter URLs, internal search, print pages, tag pages, service pagination — and confirms which of them are actually sitting in the index rather than just existing technically. It cross-checks what's already blocked in robots.txt or noindex against what's still open with no good reason.
What you get
A report in chat: a table of junk type, example URLs, roughly how much of the index they occupy, and the right fix for each — clean-param or canonical for parameters, noindex for internal search, careful handling for pagination only where it truly duplicates content. Genuinely ambiguous cases are flagged "check manually" rather than recommended for closure. Run this whenever the index looks much larger than the number of pages that actually matter, after a CMS or catalog-filter change, or as a recurring quarterly check — junk in the index rarely clears on its own.