/site-audit

name: memstack-seo-site-audit
description: "Use this skill when the user says 'SEO audit', 'site audit', 'check SEO', 'audit my site', 'SEO check', 'technical SEO', or is evaluating a website's search engine optimization health, meta tags, performance, or structured data. Do NOT use for keyword research or schema markup generation alone."
version: 1.0.0
license: "Proprietary, MemStack™ Pro by CW Affiliate Investments LLC. See LICENSE.txt"

🔎 Site Audit: Running comprehensive SEO audit...

*Scans every page for meta tags, heading hierarchy, broken links, image optimization, Core Web Vitals, robots/sitemap, performance, and structured data, producing a prioritized fix list.*

Activation

When this skill activates, output:

`🔎 Site Audit, Scanning site for SEO issues...`

Then execute the protocol below.

| Context | Status | |---------|--------| | User says "SEO audit" or "site audit" or "check SEO" | ACTIVE | | User says "technical SEO" or "audit my site" | ACTIVE | | Evaluating a website's overall search health | ACTIVE | | Optimizing specific meta tags (not full audit) | DORMANT: use meta-tag-optimizer | | Adding schema markup (not auditing) | DORMANT: use schema-markup | | Researching keywords (not auditing) | DORMANT: use keyword-research |

Anti-patterns

| Trap | Reality Check | |------|---------------| | "The site looks fine in the browser" | Crawlers and users see different things. Missing meta tags, broken links, and slow JS are invisible to humans. | | "We'll optimize after launch" | Technical debt compounds. Fixing 5 issues now prevents 50 issues later when Google has already indexed the problems. | | "One audit is enough" | Sites change with every deploy. Audit after major content or code changes, minimum quarterly. | | "Performance doesn't affect SEO" | Core Web Vitals are ranking signals. Slow sites rank lower regardless of content quality. | | "Robots.txt is set and forget" | One wrong rule can deindex your entire site. Review after every major structural change. |

Protocol

For how each AI search engine ranks and cites content (ChatGPT, Perplexity, Google AI Overview, Copilot, Claude, and Google traditional as baseline), see [`references/platform-ranking-factors.md`](./references/platform-ranking-factors.md). Use it to prioritize fixes, a client whose audience is primarily on Perplexity benefits most from FAQ schema and PDF-friendly pages, while a ChatGPT-first audience needs backlinks and 30-day-fresh content.

Step 0: Live site snapshot (optional, for deployed URLs)

If the site is already deployed, run the bundled stdlib-only audit script for a one-shot snapshot before walking the source-tree checks below:

python scripts/live_audit.py https://example.com

No API keys, no pip installs, no external dependencies: pure Python 3 stdlib (`urllib`, `html.parser`, `json`). It fetches the page once and checks title, meta description, heading hierarchy, image alt attributes, JSON-LD structured data, canonical URL, Open Graph tags, robots meta, viewport, `robots.txt` (including AI bot rules), `sitemap.xml`, page load time, and internal vs. external link counts. Output uses the same severity format (🔴 Critical / 🟠 High / 🟡 Medium / 🔵 Low) as the scorecard in Step 9, and the script exit code reflects severity (`0` clean, `1` high, `2` critical) so it can also gate CI.

Use this for deployed-URL audits; use Steps 1-8 below for source-tree scans of unshipped code. The two are complementary, the live script confirms what actually reaches crawlers, the grep steps catch issues before they ship.

---

Step 1: Check Meta Tags

Scan every page for required meta tags:

# Find all page files
find app/ pages/ src/ -name "*.tsx" -o -name "*.jsx" -o -name "*.html" | grep -v node_modules

# Search for meta tag patterns
grep -rn "title\|<meta\|og:title\|og:description\|og:image\|canonical" --include="*.tsx" --include="*.jsx" --include="*.html" . | grep -v node_modules

**Required meta tags per page:**

| Tag | Required | Rule | |-----|----------|------| | `<title>` | Yes | Unique per page, 50-60 chars, includes primary keyword | | `<meta name="description">` | Yes | Unique per page, 150-155 chars, includes CTA | | `<link rel="canonical">` | Yes | Points to the canonical URL (prevents duplicate content) | | `<meta property="og:title">` | Yes | Social sharing title (can differ from page title) | | `<meta property="og:description">` | Yes | Social sharing description | | `<meta property="og:image">` | Yes | Social sharing image (1200x630px recommended) | | `<meta property="og:url">` | Yes | Canonical URL | | `<meta name="robots">` | Conditional | `noindex` on admin, auth, and internal pages | | `<meta name="viewport">` | Yes | `width=device-width, initial-scale=1` |

**Score per page:**

• ✅ All required tags present and valid
• ⚠️ Missing optional tags
• ❌ Missing required tags or invalid values

Step 2: Check Heading Hierarchy

# Find heading usage across all pages
grep -rn "<h[1-6]\|<H[1-6]" --include="*.tsx" --include="*.jsx" --include="*.html" . | grep -v node_modules

**Heading rules:**

| Rule | Check | Why | |------|-------|-----| | Single H1 per page | Count H1 tags | Multiple H1s confuse crawlers about the page topic | | H1 matches page title (roughly) | Compare H1 to `<title>` | Consistency signals relevance | | No skipped levels | H1 → H2 → H3, not H1 → H3 | Broken hierarchy hurts accessibility and SEO | | Keywords in H1 and H2 | Check for target keyword | Headings are high-weight ranking signals | | No empty headings | Check for `<h2></h2>` | Empty headings are wasted signals | | Headings aren't used for styling | Check if headings are decorative | Use CSS for size, headings for structure |

Step 3: Check for Broken Links

# Find all links in the codebase
grep -rn 'href="\|href={' --include="*.tsx" --include="*.jsx" --include="*.html" --include="*.md" . | grep -v node_modules | grep -v "mailto:\|tel:\|#