Wp audit seo standards

Skill yojahny55/claude-wp-builder/skills/wp-audit-seo-standards

Rank Math SEO configuration reference, schema JSON-LD templates, meta patterns, and SEO seeding commandsFrom its SKILL.md

Install

npx -y skills add yojahny55/claude-wp-builder --skill wp-audit-seo-standards

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

6 stars6 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

23.2 KB, as published. Nobody here has run it

SEO Standards — Rank Math Reference

This skill defines the SEO configuration standards, schema templates, and seeding commands for WordPress sites using Rank Math SEO as the primary SEO plugin.

1. Rank Math Plugin Detection

// Plugin slug: seo-by-rank-math
// Detection: defined('RANK_MATH_VERSION') or class_exists('RankMath')
// WP-CLI install: $WP plugin install seo-by-rank-math --activate

Before running any Rank Math commands, verify the plugin is active:

$WP eval "
if (defined('RANK_MATH_VERSION')) {
    echo 'Rank Math v' . RANK_MATH_VERSION . ' is active.';
} else {
    echo 'Rank Math is NOT active.';
}
"

2. Rank Math Modules Reference

Module	Slug	Recommended	Notes
SEO Analysis	`seo-analysis`	Yes	On-page content scoring
Sitemap	`sitemap`	Yes	XML sitemap generation
Rich Snippets	`rich-snippet`	Yes	JSON-LD structured data
Breadcrumbs	`breadcrumbs`	Yes	Breadcrumb trail + schema
404 Monitor	`404-monitor`	Yes	Track broken links
Redirections	`redirections`	Yes	301/302 redirect manager
Local SEO	`local-seo`	Yes (business sites)	LocalBusiness schema
Image SEO	`image-seo`	Yes	Auto alt/title attributes
Instant Indexing	`instant-indexing`	Yes	IndexNow / Bing
Link Counter	`link-counter`	Yes	Internal/external link audit

Enable All Recommended Modules

$WP eval "
\$modules = get_option('rank_math_modules', []);
\$enable = ['seo-analysis','sitemap','rich-snippet','breadcrumbs','404-monitor','redirections','local-seo','image-seo','instant-indexing','link-counter'];
foreach (\$enable as \$mod) { if (!in_array(\$mod, \$modules)) { \$modules[] = \$mod; } }
update_option('rank_math_modules', \$modules);
echo 'Enabled modules: ' . implode(', ', \$modules);
"

3. Rank Math Option Keys

Option	Purpose
`rank_math_modules`	Array of active module slugs
`rank-math-options-general`	Breadcrumbs, links, image SEO, strip category base
`rank-math-options-titles`	Title templates, schema type, social profiles, noindex rules
`rank-math-options-sitemap`	Sitemap inclusions, ping settings
`rank-math-options-instant-indexing`	IndexNow configuration

Reading Options

# Get all general options
$WP eval "print_r(get_option('rank-math-options-general'));"

# Get all title options
$WP eval "print_r(get_option('rank-math-options-titles'));"

# Get active modules
$WP eval "print_r(get_option('rank_math_modules'));"

Writing Options

Rank Math stores options as serialized arrays. Always merge rather than overwrite:

$WP eval "
\$opts = (array) get_option('rank-math-options-general', []);
\$opts['breadcrumbs']          = 'on';
\$opts['strip_category_base']  = 'on';
\$opts['nofollow_external_links'] = 'on';
\$opts['new_window_external_links'] = 'on';
\$opts['add_img_alt']          = 'on';
\$opts['add_img_title']        = 'on';
update_option('rank-math-options-general', \$opts);
echo 'General options updated.';
"

4. Rank Math Post Meta Keys

Meta Key	Purpose	Example Value
`rank_math_title`	SEO title template	`%title% %sep% %sitename%`
`rank_math_description`	Meta description	Free text, 150-160 chars
`rank_math_focus_keyword`	Focus keyword(s)	`keyword1,keyword2`
`rank_math_robots`	Robots directives	`["index","follow"]`
`rank_math_canonical_url`	Canonical URL override	`https://...`
`rank_math_facebook_title`	OG title	Free text
`rank_math_facebook_description`	OG description	Free text
`rank_math_facebook_image`	OG image URL	`https://...`
`rank_math_facebook_image_id`	OG image attachment ID	`42`
`rank_math_twitter_card_type`	Twitter card type	`summary_large_image`
`rank_math_twitter_title`	Twitter title	Free text
`rank_math_twitter_description`	Twitter description	Free text
`rank_math_schema_Article`	Article schema	Serialized array
`rank_math_schema_FAQPage`	FAQ schema	Serialized array
`rank_math_breadcrumb_title`	Breadcrumb label override	Free text
`rank_math_pillar_content`	Pillar content flag	`on`

Seeding Post Meta via WP-CLI

# Set SEO meta for a single post
$WP eval "
\$post_id = 42;
update_post_meta(\$post_id, 'rank_math_title', '%title% %sep% %sitename%');
update_post_meta(\$post_id, 'rank_math_description', 'Your meta description here.');
update_post_meta(\$post_id, 'rank_math_focus_keyword', 'primary keyword');
update_post_meta(\$post_id, 'rank_math_robots', ['index','follow']);
update_post_meta(\$post_id, 'rank_math_twitter_card_type', 'summary_large_image');
echo 'SEO meta set for post ' . \$post_id;
"

Bulk-Seed Meta Descriptions from Content

$WP eval "
\$posts = get_posts(['post_type' => ['post','page'], 'posts_per_page' => -1, 'post_status' => 'publish']);
foreach (\$posts as \$p) {
    \$existing = get_post_meta(\$p->ID, 'rank_math_description', true);
    if (!empty(\$existing)) continue;
    \$text = wp_strip_all_tags(\$p->post_excerpt ?: \$p->post_content);
    \$desc = mb_substr(trim(preg_replace('/\s+/', ' ', \$text)), 0, 155);
    if (mb_strlen(\$desc) > 10) {
        update_post_meta(\$p->ID, 'rank_math_description', \$desc);
        echo 'Set description for: ' . \$p->post_title . PHP_EOL;
    }
}
"

5. Schema JSON-LD Templates

Organization

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "",
  "url": "",
  "logo": {
    "@type": "ImageObject",
    "url": ""
  },
  "sameAs": [
    "https://www.facebook.com/PROFILE",
    "https://www.instagram.com/PROFILE",
    "https://x.com/PROFILE",
    "https://www.linkedin.com/company/PROFILE"
  ]
}

LocalBusiness

{
  "@context": "https://schema.org",
  "@type": "LocalBusiness",
  "name": "",
  "url": "",
  "image": "",
  "telephone": "",
  "email": "",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "",
    "addressLocality": "",
    "addressRegion": "",
    "postalCode": "",
    "addressCountry": ""
  },
  "geo": {
    "@type": "GeoCoordinates",
    "latitude": "",
    "longitude": ""
  },
  "openingHoursSpecification": [
    {
      "@type": "OpeningHoursSpecification",
      "dayOfWeek": ["Monday","Tuesday","Wednesday","Thursday","Friday"],
      "opens": "09:00",
      "closes": "17:00"
    }
  ],
  "sameAs": []
}

FAQPage

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is the question?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "The answer to the question."
      }
    },
    {
      "@type": "Question",
      "name": "Another question?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Another answer."
      }
    }
  ]
}

BreadcrumbList

{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    {
      "@type": "ListItem",
      "position": 1,
      "name": "Home",
      "item": "https://example.com/"
    },
    {
      "@type": "ListItem",
      "position": 2,
      "name": "Category",
      "item": "https://example.com/category/"
    },
    {
      "@type": "ListItem",
      "position": 3,
      "name": "Current Page"
    }
  ]
}

Article

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "",
  "author": {
    "@type": "Person",
    "name": "",
    "url": ""
  },
  "datePublished": "2025-01-01T00:00:00+00:00",
  "dateModified": "2025-01-01T00:00:00+00:00",
  "image": {
    "@type": "ImageObject",
    "url": "",
    "width": 1200,
    "height": 630
  },
  "mainEntityOfPage": {
    "@type": "WebPage",
    "@id": "https://example.com/post-slug/"
  },
  "publisher": {
    "@type": "Organization",
    "name": "",
    "logo": {
      "@type": "ImageObject",
      "url": ""
    }
  }
}

6. Meta Title Formulas by Page Type

Page Type	Formula
Homepage	`%sitename% %sep% %sitedesc%`
Page	`%title% %sep% %sitename%`
Blog Post	`%title% %sep% %sitename%`
Category	`%term% %sep% %sitename% %page%`
Search	`Search: %searchphrase% %sep% %sitename%`
404	`Page Not Found %sep% %sitename%`
Author	`%name% %sep% %sitename%`

Available Variables

%title% — Post/page title
%sitename% — Site name from Settings > General
%sitedesc% — Site tagline
%sep% — Separator character (default: -)
%term% — Category/tag name
%searchphrase% — Current search query
%page% — Page number (blank on page 1)
%name% — Author display name
%date% — Post publication date
%excerpt% — Post excerpt

Seed Title Templates via WP-CLI

$WP eval "
\$opts = (array) get_option('rank-math-options-titles', []);
\$opts['homepage_title']    = '%sitename% %sep% %sitedesc%';
\$opts['pt_post_title']     = '%title% %sep% %sitename%';
\$opts['pt_page_title']     = '%title% %sep% %sitename%';
\$opts['tax_category_title'] = '%term% %sep% %sitename% %page%';
\$opts['search_title']      = 'Search: %searchphrase% %sep% %sitename%';
\$opts['404_title']         = 'Page Not Found %sep% %sitename%';
\$opts['author_archive_title'] = '%name% %sep% %sitename%';
update_option('rank-math-options-titles', \$opts);
echo 'Title templates updated.';
"

7. Meta Description Generation

Priority Order

Manual — rank_math_description post meta (highest priority)
Excerpt — Post excerpt if set
First paragraph — First <p> block in post content
Trimmed content — First 155 characters of stripped content

Target length: 150–160 characters.

Auto-Generate Description Function

function prefix_auto_meta_description( $post_id ) {
    $existing = get_post_meta( $post_id, 'rank_math_description', true );
    if ( ! empty( $existing ) ) {
        return $existing;
    }

    $post = get_post( $post_id );
    if ( ! $post ) {
        return '';
    }

    // Try excerpt first
    if ( ! empty( $post->post_excerpt ) ) {
        $text = $post->post_excerpt;
    } else {
        // Try first paragraph
        $content = $post->post_content;
        if ( preg_match( '/<p[^>]*>(.*?)<\/p>/is', $content, $matches ) ) {
            $text = $matches[1];
        } else {
            $text = $content;
        }
    }

    $text = wp_strip_all_tags( $text );
    $text = trim( preg_replace( '/\s+/', ' ', $text ) );
    $desc = mb_substr( $text, 0, 155 );

    // Avoid cutting mid-word
    if ( mb_strlen( $text ) > 155 ) {
        $desc = mb_substr( $desc, 0, mb_strrpos( $desc, ' ' ) );
        $desc .= '...';
    }

    return $desc;
}

Bulk Seed via WP-CLI

$WP eval "
\$posts = get_posts(['post_type' => ['post','page'], 'posts_per_page' => -1, 'post_status' => 'publish']);
\$count = 0;
foreach (\$posts as \$p) {
    \$existing = get_post_meta(\$p->ID, 'rank_math_description', true);
    if (!empty(\$existing)) continue;
    \$text = wp_strip_all_tags(\$p->post_excerpt ?: \$p->post_content);
    \$text = trim(preg_replace('/\s+/', ' ', \$text));
    \$desc = mb_substr(\$text, 0, 155);
    if (mb_strlen(\$text) > 155) {
        \$desc = mb_substr(\$desc, 0, mb_strrpos(\$desc, ' '));
    }
    if (mb_strlen(\$desc) > 10) {
        update_post_meta(\$p->ID, 'rank_math_description', \$desc);
        \$count++;
    }
}
echo \"Seeded \$count meta descriptions.\";
"

8. llms.txt Template

# {Site Name}

> {Site Description}

## Pages
- [Page Title](url): excerpt or first 20 words

## Blog Posts
- [Post Title](url): excerpt

## Contact
- Website: {home_url}

Generate llms.txt via WP-CLI

$WP eval "
\$name = get_bloginfo('name');
\$desc = get_bloginfo('description');
\$home = home_url('/');
\$out  = \"# \$name\n\n> \$desc\n\n\";

// Pages
\$out .= \"## Pages\n\";
\$pages = get_posts(['post_type' => 'page', 'posts_per_page' => -1, 'post_status' => 'publish', 'orderby' => 'menu_order', 'order' => 'ASC']);
foreach (\$pages as \$p) {
    \$url     = get_permalink(\$p->ID);
    \$excerpt = \$p->post_excerpt ?: wp_trim_words(wp_strip_all_tags(\$p->post_content), 20, '...');
    \$out    .= \"- [\$p->post_title](\$url): \$excerpt\n\";
}

// Blog posts
\$out  .= \"\n## Blog Posts\n\";
\$posts = get_posts(['post_type' => 'post', 'posts_per_page' => -1, 'post_status' => 'publish', 'orderby' => 'date', 'order' => 'DESC']);
foreach (\$posts as \$p) {
    \$url     = get_permalink(\$p->ID);
    \$excerpt = \$p->post_excerpt ?: wp_trim_words(wp_strip_all_tags(\$p->post_content), 20, '...');
    \$out    .= \"- [\$p->post_title](\$url): \$excerpt\n\";
}

// Contact
\$out .= \"\n## Contact\n\";
\$out .= \"- Website: \$home\n\";

file_put_contents(ABSPATH . 'llms.txt', \$out);
echo 'Generated llms.txt at ' . ABSPATH . 'llms.txt';
echo PHP_EOL . '---' . PHP_EOL . \$out;
"

9. robots.txt Template

# robots.txt for WordPress
User-agent: *
Allow: /
Disallow: /wp-admin/
Allow: /wp-admin/admin-ajax.php
Disallow: /wp-includes/
Disallow: /wp-content/plugins/
Disallow: /readme.html
Disallow: /license.txt
Disallow: /?s=
Disallow: /search/
Disallow: /cgi-bin/
Disallow: /trackback/

# AI Crawlers — Allow by default (confirm with user before changing)
User-agent: GPTBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: Google-Extended
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: CCBot
Allow: /

User-agent: Bytespider
Disallow: /

# Sitemap
Sitemap: {home_url}/sitemap_index.xml

Note: AI crawler policy (Allow vs Disallow) should be confirmed with the site owner before writing. The defaults above allow all major AI crawlers except Bytespider.

Generate robots.txt via WP-CLI

$WP eval "
\$home = home_url('/');
\$robots = 'User-agent: *
Allow: /
Disallow: /wp-admin/
Allow: /wp-admin/admin-ajax.php
Disallow: /wp-includes/
Disallow: /wp-content/plugins/
Disallow: /readme.html
Disallow: /license.txt
Disallow: /?s=
Disallow: /search/
Disallow: /cgi-bin/
Disallow: /trackback/

# AI Crawlers
User-agent: GPTBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: Google-Extended
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: CCBot
Allow: /

User-agent: Bytespider
Disallow: /

Sitemap: ' . \$home . 'sitemap_index.xml
';
file_put_contents(ABSPATH . 'robots.txt', \$robots);
echo 'Generated robots.txt at ' . ABSPATH . 'robots.txt';
"

10. Breadcrumb Integration Code

PHP Function

function prefix_breadcrumbs() {
    if (is_front_page()) return;
    echo '<nav class="breadcrumbs" aria-label="Breadcrumb">';
    if (function_exists('rank_math_the_breadcrumbs')) {
        rank_math_the_breadcrumbs();
    } else {
        echo '<a href="' . esc_url(home_url('/')) . '">Home</a>';
        echo ' &raquo; ';
        if (is_singular()) { the_title(); }
        elseif (is_archive()) { the_archive_title(); }
        elseif (is_search()) { echo 'Search results'; }
        elseif (is_404()) { echo 'Page Not Found'; }
    }
    echo '</nav>';
}

Breadcrumb CSS

.breadcrumbs {
    padding: 12px 0;
    font-size: 0.875rem;
    color: #6b7280;
}
.breadcrumbs a {
    color: #3b82f6;
    text-decoration: none;
}
.breadcrumbs a:hover {
    text-decoration: underline;
}
.breadcrumbs .separator {
    margin: 0 0.5rem;
    color: #9ca3af;
}

Enable Breadcrumbs in Rank Math

$WP eval "
\$opts = (array) get_option('rank-math-options-general', []);
\$opts['breadcrumbs'] = 'on';
update_option('rank-math-options-general', \$opts);
echo 'Breadcrumbs enabled.';
"

11. FAQ Detection Patterns

FAQ content can appear in several forms. Detect and convert to JSON-LD schema automatically.

Detection Sources

ACF repeater fields — field names: faq_items, faqs, faq_cards, faq
<details>/<summary> HTML — accordion-style FAQ in content
H2/H3 headings ending with ? — followed by <p> answer content

FAQ Auto-Detection and JSON-LD Generator

function prefix_detect_faq_jsonld( $post_id ) {
    $faqs = [];

    // 1. Check ACF repeater fields
    $acf_names = ['faq_items', 'faqs', 'faq_cards', 'faq'];
    foreach ( $acf_names as $field_name ) {
        if ( function_exists('have_rows') && have_rows( $field_name, $post_id ) ) {
            while ( have_rows( $field_name, $post_id ) ) {
                the_row();
                $q = get_sub_field('question') ?: get_sub_field('title');
                $a = get_sub_field('answer')   ?: get_sub_field('content') ?: get_sub_field('text');
                if ( $q && $a ) {
                    $faqs[] = [ 'q' => wp_strip_all_tags($q), 'a' => wp_strip_all_tags($a) ];
                }
            }
        }
    }

    // 2. Check <details>/<summary> in content
    $content = get_post_field('post_content', $post_id);
    if ( preg_match_all('/<summary[^>]*>(.*?)<\/summary>\s*(.*?)<\/details>/is', $content, $matches, PREG_SET_ORDER) ) {
        foreach ( $matches as $m ) {
            $q = wp_strip_all_tags( $m[1] );
            $a = wp_strip_all_tags( $m[2] );
            if ( $q && $a ) {
                $faqs[] = [ 'q' => $q, 'a' => $a ];
            }
        }
    }

    // 3. Check H2/H3 headings ending with ?
    if ( preg_match_all('/<h[23][^>]*>(.*?\?)<\/h[23]>\s*<p>(.*?)<\/p>/is', $content, $matches, PREG_SET_ORDER) ) {
        foreach ( $matches as $m ) {
            $q = wp_strip_all_tags( $m[1] );
            $a = wp_strip_all_tags( $m[2] );
            if ( $q && $a ) {
                $faqs[] = [ 'q' => $q, 'a' => $a ];
            }
        }
    }

    if ( empty( $faqs ) ) {
        return '';
    }

    // Build JSON-LD
    $schema = [
        '@context'   => 'https://schema.org',
        '@type'      => 'FAQPage',
        'mainEntity' => [],
    ];
    foreach ( $faqs as $faq ) {
        $schema['mainEntity'][] = [
            '@type'          => 'Question',
            'name'           => $faq['q'],
            'acceptedAnswer' => [
                '@type' => 'Answer',
                'text'  => $faq['a'],
            ],
        ];
    }

    return '<script type="application/ld+json">' . wp_json_encode( $schema, JSON_UNESCAPED_SLASHES | JSON_PRETTY_PRINT ) . '</script>';
}

Inject FAQ Schema into Head

add_action('wp_head', function() {
    if ( is_singular() ) {
        $jsonld = prefix_detect_faq_jsonld( get_the_ID() );
        if ( $jsonld ) {
            echo $jsonld;
        }
    }
});

12. SEO Seeding Sequence

Recommended order for bulk SEO setup on a new or existing site:

Install + activate Rank Math — $WP plugin install seo-by-rank-math --activate
Enable modules — activate all recommended modules (Section 2)
Configure general settings — breadcrumbs, link behavior, image SEO (Section 3)
Configure title templates — set formulas per page type (Section 6)
Configure sitemap — set post types and taxonomies to include
Set Organization / LocalBusiness schema — configure site-wide schema (Section 5)
Configure OG / Social defaults — set default OG image, social profiles
Enable IndexNow — activate instant indexing for Bing / Yandex
Seed per-page SEO meta — rank_math_title, rank_math_focus_keyword, robots (Section 4)
Seed meta descriptions — auto-generate from excerpt/content (Section 7)
Set image alt texts — bulk update missing alt attributes
Generate robots.txt — write optimized robots.txt (Section 9)
Generate llms.txt — write AI-readable site summary (Section 8)
Flush sitemap cache + rewrite rules — finalize:

$WP eval "
if (class_exists('RankMath')) {
    // Trigger sitemap regeneration
    delete_transient('rank_math_sitemap_cache');
    \RankMath\Sitemap\Cache::invalidate_storage();
}
"
$WP rewrite flush

Verification Checklist

After seeding, verify:

# Check title templates are set
$WP eval "print_r(get_option('rank-math-options-titles'));"

# Count posts with meta descriptions
$WP eval "
global \$wpdb;
\$count = \$wpdb->get_var(\"SELECT COUNT(*) FROM \$wpdb->postmeta WHERE meta_key = 'rank_math_description' AND meta_value != ''\");
echo \"Posts with meta descriptions: \$count\";
"

# Verify sitemap exists
$WP eval "echo home_url('/sitemap_index.xml');"

# Verify robots.txt exists
$WP eval "echo file_exists(ABSPATH . 'robots.txt') ? 'robots.txt exists' : 'robots.txt missing';"

# Verify llms.txt exists
$WP eval "echo file_exists(ABSPATH . 'llms.txt') ? 'llms.txt exists' : 'llms.txt missing';"

13. Lighthouse SEO Audit Gotchas (hard-won)

Non-obvious traps that make a fix look applied while the audit keeps failing. Verify against these before reporting an SEO finding fixed.

`link-text` matches VISIBLE innerText — not `aria-label`

The Lighthouse link-text audit (core/audits/seo/link-text.js) lowercases/trims the anchor's link.text (rendered innerText) and checks it against a blocklist: click here, here, learn more, more, read more, this, start, … (per-language). aria-label, title, and nodeLabel are ignored — adding an aria-label fixes the a11y accessible-name but does not satisfy this SEO audit.

To keep a design's non-descriptive label (e.g. a "LEARN MORE" button) and pass, append a visually-hidden descriptive suffix inside the anchor so it becomes part of innerText:

<a class="btn" href="<?php echo esc_url( $url ); ?>">
    <?php echo esc_html( $label ); // e.g. "LEARN MORE" — stays visible ?>
    <?php if ( $context ) : ?><span class="screen-reader-text"><?php
        echo esc_html( 'about ' . $context ); // e.g. "about Home Search"
    ?></span><?php endif; ?>
</a>

Now link.text = "LEARN MORE about Home Search" → not a blocklist match → audit passes; the button still visually reads "LEARN MORE".

Critical: the hidden span must use the .screen-reader-text clip pattern (position:absolute; clip: rect(...); width:1px) — NOT display:none or visibility:hidden, which are excluded from innerText and would fail again. (A .screen-reader-text rule is required in the theme CSS; see a11y standards.)
The same non-descriptive label often repeats across a reusable button component — fix the component/template part once, then re-check every page that uses it (Lighthouse audits per-URL; a homepage pass doesn't clear inner pages).
identical-links-same-purpose (a11y, weight 0) can also flag two same-text links (e.g. two "VIEW ALL") pointing to different destinations — the same hidden-suffix technique resolves it.

Meta description: let the SEO plugin own it — don't emit a theme fallback

If Rank Math (or Yoast) is active and configured, do not also echo a hardcoded <meta name="description"> from the theme wp_head. Two tags = a duplicate-description warning, and the plugin's dynamic/templated value is the one you want. A common failure mode: a theme adds a front-page meta-description fallback "to be safe" while the SEO plugin's setup wizard was incomplete → once the wizard is finished, the tag is duplicated. Remove the theme fallback; verify only one <meta name="description"> renders (curl -s <url> | grep -c 'name="description"' → 1).