Indexer

GSC Rapid Indexer

First: Read the file ~/.claude/skills/seo-references/core.md before proceeding. Apply the methodology in that file to all output.

Also read: ~/.claude/skills/seo-references/data-pull-patterns.md for MCP tool parameter reference. Dated GEO context: seo-references/gap-2026-08-31.md.

Core principle: connect Google Search Console, submit sitemaps, and submit URLs whenever pages are added or edited — that covers Google. ChatGPT Search and Copilot do not read GSC. For those surfaces, IndexNow + Bing Webmaster Tools is the index path.

Step 1: Parse Domain

Extract the domain. Resolve against MCP Routing Map. The GSC property is required for this skill's data pulls — if unavailable, warn but still generate advisory output.

Step 2: Pull Live Data

GSC (PRIMARY for this skill):

1. Call mcp__gscServer__list_sitemaps_enhanced — current sitemap status (submitted count, indexed count, errors). Shows health of current sitemap infrastructure.
2. Call mcp__gscServer__get_sitemaps — list existing sitemap URLs.
3. Call mcp__gscServer__inspect_url_enhanced for 3-5 key pages (homepage + top pages from portfolio context). Shows per-URL indexing status, last crawl date, and any issues.

Step 3: Generate Indexing Protocol

Section 1 — Current Indexing Status: Based on GSC data:

• How many pages are submitted vs. indexed?
• Which key pages are indexed vs. not?
• Any crawl errors or indexing issues?
• Last crawl dates — is Googlebot visiting regularly?

Section 2 — Robots.txt (advisory output): Generate an optimized robots.txt. Rules:

• Allow all important content pages
• Block: /admin, /api, /_next/static (framework internals), /staging, duplicate content paths
• Include Sitemap directive pointing to XML sitemap URL
• Retrieval-bot allow/deny is owned by /geo-crawl. Do not invent a second crawler table here.
• Note: Print to terminal for user to apply. This skill does not write to project files.

Section 3 — XML Sitemap Structure: Based on the topical map structure (13 pages per silo):

• Hub pages: priority 1.0, changefreq weekly
• Sub-Hub pages: priority 0.8, changefreq weekly
• Purchase-intent pages: priority 0.6, changefreq monthly
• Format: show the sitemap XML structure as advisory output

Section 4 — Two-index protocol

Google and Bing are different doors.

1. Google — URL Inspection + sitemap. The Indexing API is officially for JobPosting and BroadcastEvent. For BOFU pages, use URL Inspection. The legacy sitemap ping endpoint was retired in 2023; do not use it. Resubmit the sitemap in GSC after structural changes.
2. Bing — IndexNow. Participating engines include Bing and Yandex. Google is not a participant. After every BOFU publish or material edit:
   • Confirm an IndexNow key file is publicly reachable (https://[host]/[key].txt, file body = key).
   • POST the URL list to https://api.indexnow.org/indexnow with host, key, keyLocation, urlList.
   • Confirm the host in Bing Webmaster Tools. If the property has an AI Performance / Copilot citation report, screenshot the last 28 days into the playbook.
3. Manual URL Inspection remains the Google fallback when the API does not apply.

Do not tell the operator that IndexNow "indexes ChatGPT." It notifies Bing-shaped indexes that ChatGPT Search and Copilot have historically read. Retrieval can still miss the page.

Section 5 — Post-Publish Checklist (for every new purchase-intent page):

[] Submit URL in GSC via URL Inspection
[] POST the same URL to IndexNow (Bing path)
[] Confirm Bing Webmaster Tools has the host verified
[] Share on 2 social platforms (signal amplification — see /signal)
[] If this host is a SHIP/MAYBE source, confirm the Preferred Sources popup is on the page (see /preferred-source). Skip for tools and /blog paths.
[] Add internal link from hub page to the new page
[] Verify rendering in Mobile-Friendly Test
[] Check GSC indexing status after 48 hours
[] Optional: ChatGPT Search prompt for the page's money query after Bing has had a day

Section 6 — GSC generative-AI toggle Settings → Search generative AI (documented mid-2026) includes or excludes the property from AI Overviews, AI Mode, and generative Discover. Default is include. Record the current state. Do not flip it as an "optimization."

Step 4: Output

Follow the Output Protocol from core.md:

1. Print the full indexing protocol to terminal
2. Extract structured summary
3. Save to Graphiti with name Indexer — [domain]