seo-page

Use when one URL needs a bounded mixed SEO review across reader task, content, metadata, links, technical delivery, accessibility, and structured data; produce an evidence map and route each fix to it

SKILL.md

GitHub

One-page SEO Review

Markdown delivery contract

Read markdown-delivery-contract.md for every task. Write a substantive .md document for every material work step and register its direct destination in the work index.md. For public-site work, apply the owner-required Markdown discovery profile across the full eligible inventory; preserve native standards and immutable evidence formats.

Lane outputs: page-review.md; query-answer-map.md (combine small related steps with stable section links). Review the canonical HTML page and its substantive Markdown representation as one page owner. Preserve title/meta/H1 and user experience work, query intent, direct answers, entity facts, locale and source consistency.

Purpose

Assess a single page in its actual context and turn confirmed observations into specialist-owned next steps. This is a coordination/audit skill, not a generic rewrite, site crawl, implementation tool, or universal score.

Read references/page-review-protocol.md before a formal review. Read search-result-presentation.md when the page is expected to rank, improve CTR, or ship a complete search-result appearance.

Input gate

Collect the canonical/requested URL, page role, reader task, locale/device, raw and rendered capture where possible, key headers, current metadata/structured data, internal destinations, and authorized performance/indexing evidence. If the page or context cannot be accessed, explain the limitation and list safe evidence to collect.

Workflow

  1. Declare the page contract. What audience, task, site role, locale, intended query family, and canonical/indexability state does this URL serve?
  2. Inspect the rendered experience and delivery. Separate visible content/links/media from raw response, directives, canonicalization, rendering, and structured data. Keep observations tied to a capture and time.
  3. Map observations by lane. Route content usefulness and drafts to seo-content; delivery/crawling/rendering to seo-technical; structured data to seo-schema; media to seo-images; locale cluster issues to seo-hreflang; and sitemap/inventory questions to seo-sitemap.
  4. Complete search-result presentation when it is in scope. Check the unique title, descriptive meta description, visible main heading, canonical/indexability, site name context, and hostname favicon. A missing, duplicated, generic, or under-informative field is not dismissed because it is a basic item. Route copy to seo-content and favicon/crawl delivery to seo-technical.
  5. Distinguish defects from choices. A mismatch with an explicit page contract may be a confirmed defect; an apparent opportunity without evidence is a hypothesis. Avoid arbitrary meta-length, heading-count, word-count, keyword-density, or internal-link quotas. Judge meta descriptions by page-specific semantic completeness, not a hard character target.
  6. Create a handoff. Give each recommendation an evidence pointer, scope, owner, impact/risk statement, acceptance criterion, verification, and rollback if the change could harm users or discovery. Use seo-action-plan for an approved cross-team implementation plan.

Guardrails

  • Do not infer indexation, rankings, citations, or traffic from a page capture.
  • Do not change a page, deploy markup, submit URLs, or claim approval.
  • No overall page score. Preserve conflicting evidence instead of averaging it.
  • Treat content intent, technical eligibility, measurement, and platform-specific AI visibility as separate questions.

Formal evidence handoff

When this work needs a cross-team, approval-ready plan, package evidence-bound findings as an immutable seo-findings.json bundle using the checked-out suite contract. Let <suite-root> mean ${CLAUDE_PLUGIN_ROOT} in Claude Code. In Codex, read .seo-suite-runtime.json beside this SKILL.md when present and use its suite_root value; otherwise use the absolute repository checkout. Keep every referenced capture/source below raw/, retain declined claims and limitations, and run python "<suite-root>/scripts/validate_seo_findings.py" validate-findings <bundle>/seo-findings.json --bundle <bundle>. Send only a passing bundle to seo-action-plan; otherwise label the handoff provisional.

Output

Return the page contract, query-family ownership, search-result presentation matrix, actual proposed title/meta copy when evidence supports it, favicon/site-identity state, capture limitations, lane-by-lane evidence table, confirmed defects, provisional opportunities, routing/ownership, and a minimal next action list. For implementation, cite the required approved action-plan item.