hihol-labs/product-factory-os

SEO

Use this skill when PFO needs to improve organic search effectiveness for a product, landing page, documentation site, content backlog, or launch plan.

Inputs

• URL, repository path, page, product, or content set to evaluate.
• Target market, geography, language, ICP, and priority queries when known.
• Existing PFO artifacts when present: MARKET_BRIEF.md, ICP.md, GO_TO_MARKET.md, CONTENT_BACKLOG.md, FUNNEL_MODEL.md, VALIDATION_PLAN.md, TEST_PLAN.md.
• Analytics, Search Console, crawl, or rank data only when the user provides it or approves private data access.

Process

1. Read current PFO artifacts and identify the search surface: pages, docs, product category, ICP, language, and conversion goal.
2. Separate public-safe research from private data. Never send secrets, unreleased strategy, customer data, or proprietary analytics to external tools.
3. If a live URL or local site is available, inspect:
   • crawlability and indexability: status codes, redirects, robots.txt, sitemap, canonical tags, noindex, and duplicate URLs.
   • page metadata: title, description, headings, Open Graph/Twitter preview, language, and URL structure.
   • structured data: schema type, required fields, validation gaps, and search-result eligibility.
   • rendered UX: mobile readability, performance blockers, internal links, content discoverability, and conversion path.
4. Route through /browser-check when rendered UI, local preview, Core Web Vitals proxies, or visual/mobile checks affect SEO.
5. Route through /market-scan when keyword intent, competitor pages, recent public demand signals, or content opportunities require fresh public evidence.
6. Build an impact-ranked backlog:
   • technical fixes that unblock crawling, indexing, snippets, or speed.
   • content fixes that match search intent and strengthen entity/topic coverage.
   • internal-link and information-architecture fixes.
   • measurement tasks for Search Console, analytics, sitemap submission, and experiment tracking.
7. Update docs as needed:
   • SEO_AUDIT.md: current state, findings, priority, evidence, and verification.
   • SEO_GROWTH_GUARANTEE_GATE.md: baseline metric, target metric, measurement source, attribution window, implemented changes, exclusion factors, result decision, and next iteration when SEO growth is claimed or targeted.
   • CONTENT_BACKLOG.md: search-intent content candidates and briefs.
   • GO_TO_MARKET.md: organic acquisition channel assumptions.
   • VALIDATION_PLAN.md: SEO experiments, target metric, baseline, and decision rule.
   • .pfo/EXPERIMENT_PROGRAM.md: only for an approved fixed-metric SEO experiment loop.

Output Shape

Return:

SEO SURFACE:
PUBLIC/PRIVATE DATA BOUNDARY:
TECHNICAL SEO FINDINGS:
CONTENT AND INTENT GAPS:
STRUCTURED DATA:
INTERNAL LINKS:
MEASUREMENT:
SEO GROWTH GUARANTEE GATE:
PRIORITIZED BACKLOG:
PFO ARTIFACT IMPACT:
NEXT VALIDATION STEP:

When writing docs, append dated evidence instead of replacing prior user-authored research unless the user explicitly asks to rewrite.

Self-validation

Before final output, verify:

• Route, side-effect, and confirmation requirements match metadata.
• Required artifacts or read-only result are explicit.
• Verification, blockers, and next route are stated.

Rules

• Do not fabricate rankings, search volume, Search Console data, Lighthouse scores, crawl results, backlinks, or competitor evidence.
• Cite public sources when they influence a recommendation.
• Treat ranking and traffic claims as hypotheses unless backed by supplied analytics or verified public evidence.
• Do not optimize for vanity traffic when ICP, conversion goal, or funnel metric contradicts it.
• Prefer small, measurable SEO changes with a baseline and decision rule over broad content churn.
• Do not claim SEO growth as verified until SEO_GROWTH_GUARANTEE_GATE.md has a concrete baseline metric, target metric, measurement source, attribution window, implemented changes, exclusion factors, result decision, and next iteration.
• Do not write production robots.txt, sitemap, canonical, or schema changes unless the user asked for implementation; default output is an audit/backlog.