SEO Audit Skill

Perform a thorough, structured SEO audit of a given URL and produce a clear branded HTML report with findings, severity ratings, and actionable fix instructions.

Inputs

• URL (required): The page or site to audit. Ask if not provided.
• Client Name (required): Ask if not provided.
• City / Location (required): Ask if not provided.
• Date (required): Ask if not provided. Default to current month and year.
• Agency Name (optional): The "Prepared by" name on the cover page. Default to blank if not provided.
• Agency Tagline (optional): One-line description shown under the agency name.
• Scope (optional): Single page vs. site-wide. Default to single page unless told otherwise.

Audit Checklist

Use web_fetch to retrieve the page HTML, then inspect the source. Also fetch /robots.txt and /sitemap.xml from the root domain. Crawl 3–5 key internal pages if accessible (About, main service page, Contact, Blog). For items you cannot verify programmatically, note as "Manual check recommended".

1. Title Tag

• [ ] <title> present on every key page
• [ ] Title is unique per page (not duplicated across pages)
• [ ] Length is 50–60 characters — flag if over 60 (Google truncates)
• [ ] Format check: brand name + primary keyword + location modifier
• [ ] Title leads with brand or primary keyword — not a generic descriptor
• [ ] Flag if title is generic (e.g. "Home", "Welcome", or just the domain name)

2. Meta Descriptions

• [ ] <meta name="description"> present on the homepage
• [ ] Check all key pages — About, Services, Blog, Team, Contact, FAQ, individual service pages
• [ ] List every page that is missing a meta description (by name/URL)
• [ ] Length 130–160 characters
• [ ] Contains primary keyword and a clear call to action
• [ ] Unique per page — not duplicated

3. Heading Structure

• [ ] Exactly one <h1> tag per page — flag duplicates
• [ ] Flag pages with no H1 at all
• [ ] <h1> contains the primary keyword for that page
• [ ] Heading hierarchy is logical: H1 → H2 → H3 (no skipped levels)
• [ ] H2s and H3s use relevant secondary keywords

4. Keyword Targeting

• [ ] Primary keyword present in: <title>, <h1>, first paragraph, image alt text
• [ ] Location modifier present where relevant (city/region for local businesses)
• [ ] No keyword stuffing
• [ ] Service pages target specific keywords (e.g. "Plumber Toronto" not just "Plumber")
• [ ] Flag pages where keyword targeting is weak or generic

5. Image Optimisation

• [ ] All <img> tags have non-empty, descriptive alt attributes
• [ ] Alt text includes keyword + brand/location where natural
• [ ] No generic alt text like "image1.jpg" or empty alt=""
• [ ] Images have width and height attributes (prevents layout shift)
• [ ] Flag large unoptimised images — recommend WebP conversion

6. Internal Linking

• [ ] Key service/product pages linked from homepage and nav
• [ ] Anchor text is descriptive and keyword-rich
• [ ] Flag generic anchor text: "click here", "learn more", "read more"
• [ ] Count and list pages with non-descriptive anchor text
• [ ] No orphan pages (important pages with no internal links pointing to them)
• [ ] Internal links use HTTPS

7. Structured Data / Schema

• [ ] Check for <script type="application/ld+json"> blocks
• [ ] Identify schema types present
• [ ] For local businesses: flag if LocalBusiness schema is missing — should include name, address, phone, coordinates, opening hours, price range
• [ ] For service pages: flag missing FAQPage schema (targets "People Also Ask")
• [ ] For blogs: Article schema
• [ ] For products/services: Product + Offer schema
• [ ] Validate JSON-LD is well-formed
• [ ] Recommend: https://search.google.com/test/rich-results

8. Content Quality & Thin Content

• [ ] Word count per page — flag thin content under 300 words
• [ ] Text-to-HTML ratio: flag pages where code significantly outweighs readable content (common with Elementor, Divi, WPBakery)
• [ ] Service pages should have minimum 400–600 words with FAQs and details
• [ ] Clear call-to-action present on each key page
• [ ] Flag pages that would benefit from FAQ sections

9. Links

• [ ] No broken or malformed href values
• [ ] External links to low-authority or spammy domains (flag if suspicious)
• [ ] All links use HTTPS

10. URL & Canonicalisation

• [ ] URLs are clean, lowercase, use hyphens not underscores
• [ ] Canonical tag (<link rel="canonical">) present on all pages
• [ ] No duplicate content from query strings or pagination
• [ ] HTTPS in use site-wide
• [ ] No trailing slash inconsistency

11. Technical SEO

• [ ] robots.txt accessible — review for accidental disallows
• [ ] sitemap.xml accessible
• [ ] <meta name="robots"> — check for accidental noindex or nofollow
• [ ] <html lang="..."> attribute set correctly
• [ ] No render-blocking scripts in <head>
• [ ] Favicon present
• [ ] <meta charset> and <meta name="viewport"> present
• [ ] Open Graph tags: og:title, og:description, og:image, og:url
• [ ] Twitter Card tags present
• [ ] Note: Check Google Search Console for crawl errors (manual check — advise client)

12. Page Speed & Performance

• [ ] Flag large or unoptimised images — recommend WebP format
• [ ] Check for heavy page builder markup causing bloated HTML
• [ ] Flag excessive plugins and third-party scripts
• [ ] Recommend: compress images, reduce plugins, implement caching
• [ ] Recommend CDN integration
• [ ] Always link to: https://pagespeed.web.dev/
• [ ] Note: Core Web Vitals (LCP, CLS, FID) require manual tool check

13. Local SEO Signals

• [ ] NAP (Name, Address, Phone) consistent and present on homepage and contact page
• [ ] Google Business Profile linked or referenced
• [ ] Local keywords used in title, H1, and copy (city + service)
• [ ] Recommend local citations: Yelp, industry directories, local business listings
• [ ] Note star rating and review count if visible in schema or page
• [ ] Recommend prompting customers to leave reviews mentioning specific services and location

14. WordPress-Specific (if detected)

• [ ] SEO plugin active — Yoast SEO or Rank Math
• [ ] Plugin audit: identify unnecessary plugins adding overhead or script bloat
• Flag backup plugins if a better solution exists
• Flag unused tracking/form plugins (recommend GTM instead)
• Flag RSS feeds if not needed
• Flag unused activity log plugins
• [ ] Default ?p=123 permalinks in use? (should use post name)
• [ ] Check for duplicate meta tags from theme + plugin conflicts
• [ ] xmlrpc.php exposure (crawl budget + security concern)

15. Shopify-Specific (if detected)

• [ ] Collection and product pages have unique meta descriptions
• [ ] /collections/all — check if indexed (often thin content, consider noindex)
• [ ] Canonical tags on paginated collection pages (?page=2)
• [ ] Product image alt text populated
• [ ] Duplicate content between /products/X and /collections/Y/products/X

Severity Ratings

Output Format

Always produce the report as a downloadable PDF file using the following process:

Step 1 — Build the HTML source

Create a full HTML file (/home/claude/seo-audit-report.html) with these two pages:

Cover page:

• "SEO AUDIT REPORT" heading on dark teal (#005F7F) background
• Cyan (#3AC1CD) diagonal accent strip below header
• Client name, website URL, city/location, date
• "Prepared by" footer: Agency Name and Agency Tagline if provided, blank if not

Report page:

• Score ring (SVG circle) out of 100
• Stat grid: Critical / High / Medium / Passing counts
• Summary paragraph (2–3 sentences, most critical issues upfront)
• Findings grouped by severity with colour-coded badges
• Each finding: issue name, what was found, why it matters, exact fix instruction
• List affected pages by name/URL where multiple pages share the same issue
• Recommended next steps (max 5–6, numbered) on dark teal background
• Tools row at bottom: PageSpeed Insights, Google Search Console, Rich Results Test, Screaming Frog

Design Tokens

• Brand colours: #005F7F dark teal, #3AC1CD cyan, #EA661B orange
• Font: Montserrat via Google Fonts (@import in <style>)
• Bold 700 headings, 500 subheadings, 400 body, 11px minimum font size
• A4 page size (@page { size: A4; margin: 0; })
• Use page-break-after: always on cover page
• Use page-break-inside: avoid on each finding card

Report Sections (always include all)

1. Technical SEO Findings
2. Content & Keyword Opportunities
3. Website Speed & Performance
4. Local SEO Signals
5. Plugin Audit (if WordPress/Shopify detected)
6. Recommended Next Steps

Step 2 — Convert HTML to PDF

import subprocess
subprocess.run([
    'python3', '-c',
    '''
import sys
try:
    from weasyprint import HTML
    HTML(filename="/home/claude/seo-audit-report.html").write_pdf("/mnt/user-data/outputs/seo-audit-report.pdf")
    print("PDF generated with WeasyPrint")
except ImportError:
    import subprocess
    result = subprocess.run(
        ["chromium", "--headless", "--no-sandbox", "--print-to-pdf=/mnt/user-data/outputs/seo-audit-report.pdf",
         "--print-to-pdf-no-header", "/home/claude/seo-audit-report.html"],
        capture_output=True, text=True
    )
    if result.returncode != 0:
        subprocess.run(
            ["wkhtmltopdf", "--page-size", "A4", "--margin-top", "0",
             "--margin-bottom", "0", "--margin-left", "0", "--margin-right", "0",
             "/home/claude/seo-audit-report.html", "/mnt/user-data/outputs/seo-audit-report.pdf"],
            check=True
        )
        print("PDF generated with wkhtmltopdf")
'''
], check=True)

Name the output file: seo-audit-[client-slug]-[month]-[year].pdf Example: seo-audit-halalgrocery-may-2026.pdf

Then use present_files to deliver the PDF to the user.

Execution Steps

1. Ask for URL, client name, city, and date if not provided.
2. Use web_fetch to retrieve the full HTML of the homepage.
3. Fetch /robots.txt and /sitemap.xml.
4. Crawl 3–5 key internal pages if accessible.
5. Work through all 15 checklist sections systematically.
6. Detect CMS from HTML clues (WordPress: wp-content, wp-json; Shopify: cdn.shopify.com).
7. Apply CMS-specific checks.
8. Compile all findings with severity ratings.
9. Build the branded HTML file at /home/claude/seo-audit-report.html.
10. Install weasyprint if not available: pip install weasyprint --break-system-packages -q
11. Convert HTML to PDF and save to /mnt/user-data/outputs/.
12. Use present_files to deliver the PDF download to the user.

Notes

• Always be specific — quote the exact tag, value, or page name that has a problem.
• List every affected page by name/URL where multiple pages share the same issue.
• Keep tone professional and plain-English — the client may not be technical.
• If a page blocks fetching (403, JS-rendered SPA), note this and advise browser-based tools.
• Always include actionable fix instructions, not just observations.