We automated our entire SEO & GEO audit process

#!/usr/bin/env python3
"""
SEO Audit: Site Crawler
Checks all sitemap URLs and crawls pages for broken internal links.
Outputs machine-verified CSV data — no AI interpretation.

Usage: python3 crawl.py <domain> <sitemap_url> <output_dir>
"""
import xml.etree.ElementTree as ET
import urllib.request
import urllib.error
import ssl, sys, csv, time, re, json, os
from html.parser import HTMLParser
from urllib.parse import urlparse, urljoin

class LinkExtractor(HTMLParser):
    def __init__(self, base_url):
        super().__init__()
        self.base_url = base_url
        self.links = []
    def handle_starttag(self, tag, attrs):
        if tag == 'a':
            for name, value in attrs:
                if name == 'href' and value:
                    self.links.append(urljoin(self.base_url, value))

def fetch_url(url, timeout=15):
    ctx = ssl.create_default_context()
    req = urllib.request.Request(url, headers={
        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) ...'
    })
    try:
        resp = urllib.request.urlopen(req, timeout=timeout, context=ctx)
        return resp.getcode(), resp.read().decode('utf-8', errors='replace'), resp.geturl()
    except urllib.error.HTTPError as e:
        return e.code, '', url
    except Exception as e:
        return 0, '', str(e)

# ... parses sitemap XML recursively
# ... checks every URL for 200 status
# ... crawls all pages for internal links
# ... verifies each internal link
# ... outputs CSV with broken_url, status, found_on

---
name: seo-audit-meta
description: Machine-verified meta tag extraction. Extracts title,
  description, H1, H2, word count, canonical, OG tags from every page.
disable-model-invocation: true
allowed-tools: Bash
argument-hint: "[domain] [sitemap-url] [output-dir]"
---

# SEO Audit: Meta Tag & Content Extraction

python3 ${CLAUDE_SKILL_DIR}/scripts/meta.py $ARGUMENTS

Output:
- {domain}_meta.csv — All meta data per URL
- {domain}_meta_issues.csv — Auto-flagged issues

---
name: seo-audit-schema
description: Machine-verified JSON-LD schema extraction from every page.
disable-model-invocation: true
allowed-tools: Bash
argument-hint: "[domain] [sitemap-url] [output-dir]"
---

Output:
- {domain}_schema.csv — Every JSON-LD block with type and preview
- {domain}_schema_pages.csv — Per-page schema count
- {domain}_schema_summary.json — Type counts and statistics

---
name: seo-audit-images
description: Machine-verified image alt text audit.
disable-model-invocation: true
allowed-tools: Bash
argument-hint: "[domain] [sitemap-url] [output-dir]"
---

Output:
- {domain}_images.csv — Every image with alt text status
- {domain}_images_summary.csv — Per-page summary
- {domain}_images_stats.json — Overall statistics

---
name: seo-audit-hreflang
description: Machine-verified hreflang extraction and verification.
disable-model-invocation: true
allowed-tools: Bash
argument-hint: "[domain] [sitemap-url] [output-dir]"
---

Output:
- {domain}_hreflang.csv — Every hreflang tag with alternate URL status
- {domain}_hreflang_summary.json — Summary statistics

---
name: seo-audit-ai-crawlers
description: Tests AI search crawler access
  (GPTBot, ClaudeBot, PerplexityBot).
disable-model-invocation: true
allowed-tools: Bash
argument-hint: "[url] [output-dir]"
---

Tests 9 user agents + robots.txt + llms.txt
Output:
- {domain}_ai_crawlers.csv — Status code per crawler
- {domain}_ai_crawlers_summary.json — Blocked crawlers list

---
name: seo-audit-gsc
description: Pull Google Search Console data for SEO audit.
disable-model-invocation: true
argument-hint: "[gsc-property] [days]"
---

# SEO Audit: Google Search Console Data Pull

Uses GSC MCP tools:
- mcp__gsc__get_performance_overview
- mcp__gsc__get_search_analytics

Dimensions: query, page, device, country
Property format: sc-domain:example.com

---
name: seo-audit-ga
description: Pull Google Analytics 4 data for SEO audit.
disable-model-invocation: true
argument-hint: "[ga4-property-id] [days]"
---

Uses: mcp__analytics-mcp__run_report

Reports:
- Channel: sessions, users, pageviews, bounce rate
- Landing pages: top 20 by sessions
- Device: mobile vs desktop vs tablet

---
name: seo-audit-performance
description: Run Lighthouse audits and performance traces.
disable-model-invocation: true
argument-hint: "[url1] [url2] [url3] ..."
---

Uses Chrome DevTools MCP:
- mcp__chrome-devtools__lighthouse_audit
- mcp__chrome-devtools__performance_start_trace
- mcp__chrome-devtools__performance_analyze_insight

Flags: LCP > 2500ms, INP > 200ms, CLS > 0.1

---
name: seo-audit-keywords
description: Import and analyze keyword ranking data from CSV exports.
disable-model-invocation: true
allowed-tools: Bash, Read, Write, Glob, Grep
argument-hint: "[csv-file-or-directory]"
---

Identifies:
- Keywords at positions 4-10 (push to top 3)
- Keywords at positions 11-20 (push to page 1)
- Keywords at 101+ with high volume (not ranking)
- Wrong page ranking for a keyword