seo-optimizer

When to use

Trigger this skill when the request includes any of:

• "SEO review this draft before it goes live"
• "Do keyword research for [topic]"
• "Optimize this landing page"
• "What should the meta tags be for the new case study?"
• A finished draft plus a request to "make sure this ranks"

Do not use for paid search copy (different playbook), technical SEO audits of a whole domain (that is a separate engagement), or link-building outreach.

Inputs

Required:

• Page URL or draft content — the thing being optimized.
• Primary goal — one of: rank for a specific query, capture organic traffic for a topic cluster, or improve CTR on an existing page.

Optional:

• Seed keyword — if known. If not, the skill will derive one from the content.
• Target audience intent — helps disambiguate between a "how to" and "best X tools" angle.
• Competitor URLs already ranking for the query.

Outputs

A single SEO brief containing:

1. Keyword map — seed keyword, 8-15 related queries grouped by intent (informational, navigational, commercial, transactional).
2. On-page recommendations:
   • Title tag (≤60 chars).
   • Meta description (≤155 chars).
   • H1 (exactly one, matches intent).
   • H2/H3 outline with keywords in natural placement.
   • Internal links (minimum 3, to relevant cluster pages).
   • External links (minimum 1 to a high-authority source).
   • Image alt text suggestions for every image.
3. Schema.org markup — the JSON-LD block to embed (Article, FAQPage, HowTo, Product, or BreadcrumbList as appropriate).
4. Core Web Vitals checklist — what to verify before publish.
5. Gap note — 2-3 bullets on what the current draft is missing vs the top-ranking competitors.

Tool dependencies

• WebSearch / web-fetch (required) — for SERP checks, competitor pulls, and Google auto-suggest.
• Optional: Google Search Console MCP (if available) for existing-page CTR data.

Procedure

1. Keyword research

1. Start with the seed. If not provided, extract it from the page's thesis. The seed should be 2-4 words.
2. Pull related queries. Use WebSearch for:
   • Google auto-suggest (seed + a-z) for long-tail variants.
   • "People also ask" box for the seed.
   • Top 10 SERP results — scan their H1s and H2s for recurring phrases.
3. Classify intent for each query:
   • Informational: "what is X," "how does X work," "X explained" — user wants to learn.
   • Navigational: "X brand name," "X login," "X docs" — user wants a specific site.
   • Commercial: "best X," "X vs Y," "X alternatives," "X reviews" — user is comparing.
   • Transactional: "buy X," "X pricing," "X free trial," "X demo" — user is ready to act.
4. Filter to relevance. Drop queries that do not match the page's goal. A case study should not try to rank for transactional queries.