← ClaudeAtlas

keyword-clusteringlisted

Group keywords into intent-led page opportunities using metrics, live SERP overlap, and existing relevant pages.
SamurAIGPT/ai-seo-agent · ★ 1 · Data & Documents · score 78
Install: claude install-skill SamurAIGPT/ai-seo-agent
# Keyword Clustering ## Mission Map a keyword set to the smallest sensible set of pages. Use shared ranking URLs and SERP composition as evidence of intent; semantic similarity alone is not enough. ## Required inputs - A keyword list or a source report from Keyword Research. - Target domain when existing page mapping is requested. - Country, language, and device. - Optional: maximum clusters, page types, and business priorities. ## Workflow 1. Read the source keyword list and preserve its metrics and provenance. 2. If a direct Search Console connection is available, query `gsc.search_analytics` with `query` and `page` dimensions. Use the returned query-to-page rows as first-party evidence for existing coverage and cannibalization; do not treat missing rows as proof of no impressions. 3. Normalize terms for comparison but retain the original text and spelling. 4. Call seo.keyword_overview for explicit terms when metrics are absent or need to be made comparable. 5. Group obvious variants provisionally by topic, modifier, and funnel intent. 6. Select representative terms from each provisional group. Call seo.google_serp for those terms using the same market, language, device, and depth. Do not call every term unless the user asks for exhaustive evidence. 7. Compare the top organic URLs and result types. Merge terms when the same pages and intent dominate; split terms when the SERP leaders or user intent differ materially. 8. Call seo.relevant_pages