/skill

SEO

Original skill by ECC. Credit to ECC for the core approach and intent.

Improve search reach through sound site setup, fast pages, and useful content. Do not use tricks.

When to use this skill

Use this skill when the user asks you to:

• Check crawl or index problems
• Fix canonical tags or redirects
• Improve titles, descriptions, or headings
• Add or check structured data
• Improve Core Web Vitals
• Map search terms to pages
• Plan internal links
• Edit a sitemap or robots.txt
• Find pages that target the same search need

Work steps

1. Read the real page and related files.
2. Learn the page goal and target user.
3. Check crawl and index blocks.
4. Check the main URL, canonical tag, and redirects.
5. Check page speed and mobile use.
6. Check the title, description, headings, and body.
7. Check structured data.
8. Check internal links and the sitemap.
9. List each issue with its place, impact, and fix.
10. Make changes only when the user asks for them.
11. Test each change after it is made.

Fix high-impact site problems before small content edits.

Core rules

• Give each page one clear main search goal.
• Write for people first.
• Use mobile pages as the main view.
• Tie each fix to a page, file, or URL.
• Do not promise a rank or traffic gain.
• Do not invent search volume, rank, or site data.
• State when a claim needs data from search or speed tools.
• Keep pages that serve different user needs.
• Merge or redirect pages that serve the same need and add no clear value.
• Keep current URLs unless a change has a clear gain.
• If a URL must change, add a direct permanent redirect and update links to it.

Technical checks

Crawl access

• Make sure important pages are not blocked by robots.txt.
• Do not treat a robots.txt block as a safe way to remove a page from search.
• Check that important links use real links with valid URLs.
• Keep key pages within a few useful clicks of a main page.
• Avoid redirect chains. Point links at the final URL.
• Find broken links, redirect loops, and soft 404 pages.
• Check pages that need JavaScript to show key text or links.

Index access

• Make sure key pages do not have an unwanted noindex.
• Check both HTML meta tags and HTTP headers for index rules.
• Use one clear URL form for HTTP or HTTPS, host name, slash style, and letter case.
• Use an absolute canonical URL.
• Point the canonical tag to the best indexable page.
• Do not create canonical loops or mixed signals.
• Keep only canonical, public, successful URLs in XML sitemaps.
• Remove redirected, missing, blocked, and noindex URLs from sitemaps.
• Split large sitemaps when needed.
• Check duplicate pages caused by filters, sort rules, tags, print views, or URL values.
• For language or region pages, use valid and matching hreflang links.
• Make hreflang links return the link on each matching page.
• Use x-default only when there is a true default page.

Redirects and errors

• Use a permanent redirect for a lasting URL move.
• Use a temporary redirect only for a short change.
• Redirect old pages to the closest useful match.
• Do not send many unrelated old pages to the home page.
• Keep true missing pages as 404 or 410.
• Do not hide error pages behind a 200 status.

Core Web Vitals

Aim for these user test results at the 75th percentile:

• LCP: 2.5 seconds or less
• INP: 200 milliseconds or less
• CLS: 0.1 or less

Common fixes:

• Make the main image small and fast.
• Preload the main image only when it is known and needed.
• Do not lazy-load the main image above the fold.
• Set image width and height.
• Save space for ads, embeds, and late content.
• Cut unused JavaScript and CSS.
• Split long tasks.
• Delay work that is not needed at page load.
• Use good cache rules.
• Keep server response time low.
• Use field data when it exists. Lab tests help find causes, but they do not show all real users.

Page content checks

Title

• Make each important page title unique.
• Put the main topic near the start when it reads well.
• State the page value in clear words.
• Use about 50 to 60 characters as a guide, not a hard rule.
• Avoid repeated words and keyword lists.
• Keep the title in line with the page content.

Meta description

• Describe the page in a true and useful way.
• Add the main topic in natural words.
• Give the user a clear reason to visit.
• Use about 120 to 160 characters as a guide, not a hard rule.
• Make key page descriptions unique.
• Know that search sites may show different text.

Headings

• Use one clear main heading in most cases.
• Make the main heading match the page goal.
• Use H2 and H3 for real sections and sub-sections.
• Do not skip levels just to change text size.
• Do not use headings as style hooks only.

Main content

• Answer the main need early.
• Add facts, examples, steps, or proof that help the user.
• Remove text that repeats the same point.
• Show who wrote or checked the page when trust matters.
• Keep names, dates, prices, and claims up to date.
• Do not add text only to make a page longer.

Structured data

Add structured data only when it matches content that users can see.

Common types:

• Home page: Organization or a fitting local business type
• News or blog page: Article or BlogPosting
• Product page: Product with a valid Offer when price and stock are shown
• Inner page: BreadcrumbList
• Question page: FAQPage only when the page and search rules allow it

Rules:

• Use the most exact type that fits.
• Include all required fields.
• Use real names, URLs, dates, prices, and stock states.
• Keep markup in step with the page.
• Do not mark hidden or false content.
• Do not expect rich search results just because markup is valid.
• Check JSON-LD syntax after each edit.

Example:

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "How to Care for a Cast-Iron Pan",
  "author": {
    "@type": "Person",
    "name": "Sam Lee"
  },
  "publisher": {
    "@type": "Organization",
    "name": "Kitchen Guide"
  },
  "datePublished": "2026-02-10",
  "dateModified": "2026-03-04"
}

Keyword mapping

1. Name the user need behind each search.
2. Group terms that share the same need.
3. Check which current page best meets each need.
4. Map one main topic group to one URL.
5. Add close terms only when they fit the same page.
6. Find two or more pages that fight for the same need.
7. Pick one main page, then merge, redirect, or narrow the other pages.
8. Create a new page only when no good page exists.

Do not make up keyword data. If no trusted data is given, label the map as a topic map, not keyword research.

Use this table:

Internal links

• Link from useful, trusted pages to pages that need support.
• Use link text that tells users what they will open.
• Vary link text only when the words stay clear and true.
• Add links where they help the reader.
• Link new pages to related old pages.
• Link related old pages back to the new page.
• Fix broken links and links that pass through redirects.
• Do not add large blocks of weak links to every page.

robots.txt and sitemap safety

Before changing either file:

• Read the full current file.
• Check every rule that may affect the target path.
• Keep rules for private areas, search pages, carts, and admin pages when they are still needed.
• Do not block CSS or JavaScript files needed to show the page.
• Do not place secret paths in robots.txt.
• Keep sitemap URLs absolute.
• Test patterns against sample URLs.
• State which URLs will change before making a wide rule.

Audit output

Sort findings by impact:

• HIGH: Blocks crawling, indexing, use, or a main business page
• MEDIUM: Hurts page meaning, speed, links, or search display
• LOW: Small cleanup with limited likely impact

Use this form:

[HIGH] Duplicate titles on product pages
Place: src/routes/products/[slug].tsx
Pages: /products/red-mug and /products/blue-mug
Issue: Both pages use the same default title. Search tools may not know which page fits each product.
Fix: Build a unique title from the product name and main group.
Check: Open both pages and confirm that each HTML title is unique.

Concrete usage example

User request:

Audit the SEO for my shop pages and fix the top issues.

Process:

1. Read the shop routes, page templates, robots.txt, and sitemap.
2. Check status codes, index rules, canonicals, titles, headings, and product markup.
3. Compare similar product URLs for duplicate search goals.
4. List findings by impact.
5. Fix the high-impact issues in the named files.
6. test the changed pages and JSON-LD.
7. Report what changed and what still needs real user or search data.

Good result:

[HIGH] Product pages point to the shop home as canonical
Place: src/pages/products/[slug].tsx
Impact: Search tools may drop the product URLs from the index.
Fix: Set each product canonical to its own final public URL.
Check: Render two products and confirm that each has one self-pointing canonical.

Common mistakes

Related skills

• seo-specialist
• frontend-patterns
• brand-voice
• market-research