seo-google-leak-deep-audit

name	seo-google-leak-deep-audit
description	SEO audit of a website through the lens of the Google Content Warehouse API leak and the logic of Mustang -> Twiddlers -> NavBoost. Checks 12 structural errors, builds an evidence-based report, and generates an action plan across 7 strategic growth points.
metadata	{"triggers":["Google leak","Content Warehouse API","Mustang","Twiddlers","NavBoost","siteAuthority","NSR","siteRadius","siteFocusScore","contentEffort","clutterScore","anchorMismatchDemotion"]}

SEO Google Leak Skill

Using seos-cli in this skill

Adoption decision: seos-cli is optional and not used for the core leak audit by default. This skill's own scripts handle bounded crawling, GSC/GA4 imports, RDAP, visual checks, proxy metric extraction, and leak-specific report generation.

Use seos-cli only as supplemental evidence when the audit specifically needs project-native scrape, SERP, or Trends exports outside the core pipeline:

seos-cli config check
seos-cli scrape --urls https://example.com/page --format json --out outputs/leak-audit-scrape.json
seos-cli serp --keywords "target query" --format xlsx --out outputs/leak-audit-serp.xlsx
seos-cli trends --keywords "target keyword, competitor keyword" --format xlsx --out outputs/leak-audit-trends.xlsx

Do not use seos-cli instead of scripts/leak_audit.py for crawl limits, proxy checks, GSC/GA4 CSV imports, RDAP host-age data, Playwright visual checks, or final action-plan generation. Do not claim seos-cli computes Google leak metrics.

Before running (mandatory questions)

1. Target URL and crawl limit (--max-pages, default 300).
2. Target language/languages (or auto-detection).
3. Are there any CSV exports available (which ones and how to get them):
   • GSC Pages
     • Where: Google Search Console -> Performance -> Search results.
     • How: Search type: Web -> set period (recommended Last 3 months, minimum 28 days) -> Pages tab -> Export -> Download CSV.
     • What should be in CSV: URL/Page + metrics Clicks, Impressions, CTR, Position.
     • Recommended file name: gsc_pages.csv.
   • GSC Queries
     • Where: same GSC report Performance -> Search results.
     • How: keep the same period and filters, switch to Queries tab -> Export -> Download CSV.
     • What should be in CSV: Query + (preferably) Clicks, Impressions, CTR, Position.
     • Recommended file name: gsc_queries.csv.
   • GA4
     • Where: Google Analytics 4 -> Reports -> Engagement -> Pages and screens.
     • How: set the same period as in GSC -> Primary dimension: Page path + query string and screen class (or Page location) -> Share this report / Export file -> CSV.
     • What should be in CSV: URL/Page path (pagePath/pageLocation) + metrics Views, Users, Engagement rate, Average engagement time, Bounce rate (if available).
     • Recommended file name: ga4_pages.csv.
   • Before running, check that all CSVs are in UTF-8 and passed to the CLI like this:
     • --gsc-pages "D:/data/gsc_pages.csv" --gsc-queries "D:/data/gsc_queries.csv" --ga4 "D:/data/ga4_pages.csv"
4. Is the Playwright visual mode needed (auto|true|false).
5. Include noindex pages in the analysis? (--include-noindex, default false — noindex pages are not part of the 12 checks, but always appear in a separate block in the report to detect accidental noindex).

What the skill does

• Crawls the site with a safe, limited traversal.
• Extracts proxy signals for Mustang, Twiddlers, NavBoost.
• Computes 12 checks from the methodology and provides evidence URLs.
• If GSC/GA4 are present, prioritizes NavBoost-risk pages.
• Generates:
  • reports/LEAK-AUDIT-REPORT.md
  • reports/LEAK-ACTION-PLAN.md
  • cache/crawl.jsonl
  • cache/analysis.json

How to run

cd "D:/Projects/auto-seo-keyword-planner/skills/seo-google-leak-deep-audit"
python scripts/leak_audit.py https://example.com ^
  --max-pages 300 ^
  --concurrency 5 ^
  --delay-ms 200 ^
  --respect-robots true ^
  --sitemap "https://example.com/sitemap.xml" ^
  --sitemap "D:/data/sitemap.xml" ^
  --sitemap-page "https://example.com/site-map" ^
  --use-playwright auto ^
  --include-noindex false ^
  --out-dir reports ^
  --cache-dir cache

With external data:

python scripts/leak_audit.py https://example.com ^
  --sitemap "https://example.com/sitemap.xml" ^
  --sitemap-page "https://example.com/site-map" ^
  --gsc-pages "D:/data/gsc_pages.csv" ^
  --gsc-queries "D:/data/gsc_queries.csv" ^
  --ga4 "D:/data/ga4_pages.csv"

Quick smoke-check

python scripts/leak_audit.py https://example.com --max-pages 30

Limitations v1

• Proxy metrics are used, not real internal Google signals.
• Visual checks require Playwright to be installed.
• There are no integrations with Ahrefs/Semrush API in v1.

Metric Coverage Expansion Inputs

Additional optional exports are now supported for richer NavBoost and metric coverage:

• --gsc-timeseries "D:/data/gsc_timeseries.csv"
  • CSV headers: Date,Page,Clicks,Impressions,CTR,Position
  • Export from GSC Performance report with Date + Page dimensions.
• --ga4-traffic "D:/data/ga4_traffic.csv"
  • CSV headers: Page,Sessions,Direct Sessions
  • Export from GA4 report filtered by page URL/path.
• --ga4-scroll "D:/data/ga4_scroll.csv"
  • CSV headers: Page,Scroll Events,Users
  • Export from GA4 events report with scroll event counts.

Example:

python scripts/leak_audit.py https://example.com ^
  --gsc-pages "D:/data/gsc_pages.csv" ^
  --gsc-queries "D:/data/gsc_queries.csv" ^
  --ga4 "D:/data/ga4_pages.csv" ^
  --gsc-timeseries "D:/data/gsc_timeseries.csv" ^
  --ga4-traffic "D:/data/ga4_traffic.csv" ^
  --ga4-scroll "D:/data/ga4_scroll.csv"

Report outputs now include:

• Data Sources Used
• Metric Coverage Matrix
• Authority Graph Panel
• NavBoost Panel (COEC + demotion candidates + direct/scroll proxies)

RDAP / hostAge

• The hostAge metric is now determined automatically by domain from the input site_url.
• Source: RDAP (https://rdap.org/domain/<domain>), without hardcoding the domain in the skill or scripts.
• In cache/external.json, an rdap_whois block is saved if the registration event was successfully obtained.
• In analysis.json -> proxy_metrics, the following appear:
  • hostAge_proxy_years
  • hostAge_proxy_days
• In the Metric Coverage Matrix, the hostAge metric gets the status implemented if RDAP returned a registration date.
• If RDAP is unavailable or did not return a registration event, hostAge remains in the missing-data status.