Search indexability optimization

Overview

Use this skill to diagnose search discovery and indexability foundations before proposing content or generative-visibility interventions. It separates observed blockers and degradations from opportunities and unavailable evidence.

The skill does not operate a crawler, change directives, deploy redirects, guarantee indexing, or infer search-engine state from a single page inspection.

When to Use

• Search systems may not discover, access, render, index, or interpret important surfaces.
• Duplicate, migrated, localized, filtered, dynamic, or legacy URLs create conflicting signals.
• Navigation, hierarchy, internal links, or orphaned content affect discoverability.
• Structured data or meaningful text may not match visible content.
• Search performance evidence must be connected to technical hypotheses.

When NOT to Use

• The dominant issue is inconsistent people, brands, methods, products, or claims; use knowledge-entity-representation.
• The task is only to write or restructure content without a search-access decision.
• The requester wants rank guarantees, a production crawl, monitoring, or automated changes.
• The review has no authorized property scope or evidence source.

Core Moves

1. Bound the indexability intent. Name property, environments, audiences, important surfaces, expected access, exclusions, decision owner, and available evidence.
2. Inspect discovery and access signals. Review available URL inventories, responses, directives, sitemaps, rendering, authentication, resources, and crawler observations.
3. Evaluate interpretation paths. Review canonicalization, duplication, hierarchy, navigation, internal relationships, accessible text, metadata, and structured representation.
4. Classify evidence. Separate blockers, degradations, risks, opportunities, and unknowns; distinguish observed facts from inferences and hypotheses.
5. Prioritize verification and handoff. Connect each finding to consequence, bounded remediation, owner, and an observable recheck without changing the property.

Optional Modules

• Crawl access review — Inspect directives, response behavior, authentication, resources, rendering dependencies, and crawler access.
• Indexability and canonicalization review — Inspect canonicals, redirects, noindex, duplication, pagination, locale, versions, filters, and conflicting URLs.
• Information architecture review — Inspect hierarchy, navigation, internal links, discoverability depth, semantic grouping, and orphaned surfaces.
• Structured-data consistency review — Compare structured data with visible content, supported types, entities, and relationships.
• Content accessibility review — Check whether meaningful content is available as accessible text rather than only in images, interaction, or client-only states.
• Search performance baseline — Use authorized webmaster or analytics evidence to compare technical hypotheses with observed discovery and search behavior.

Activation Triggers

• Activate crawl access review when access is blocked, dynamic, authenticated, inconsistent, or environment-dependent.
• Activate indexability and canonicalization review for duplicates, migrations, locales, filters, pagination, multiple versions, or legacy URLs.
• Activate information architecture review when several offerings, audiences, content types, or future expansion paths exist.
• Activate structured-data consistency review when structured data exists or rich-result and entity interpretation matter.
• Activate content accessibility review for JavaScript-heavy, visual, multimedia, or interaction-dependent surfaces.
• Activate search performance baseline only when authorized performance evidence exists.

Expected Output

search_indexability_review:
  scope:
    property:
    environments: []
    important_surfaces: []
    indexability_intent:
    observed_at:
  findings:
    - finding_id:
      skill: search-indexability-optimization
      module:
      target:
      finding_type: blocker | degradation | risk | opportunity | unknown
      evidence:
        source:
        observed_at:
        method:
        raw_observation:
      confidence: high | medium | low | unavailable
      fact:
      inference:
      hypothesis:
      impact:
      recommendation:
      verification:
      handoff:
  remediation_sequence: []
  unavailable_evidence: []
  limitations: []

Verification

• Scope, environment, important surfaces, expected access, and intended exclusions are explicit.
• Access, crawl, render, index, and performance claims identify their evidence and date.
• Canonical, redirect, directive, structured-data, and architecture findings describe the observed target rather than a remembered provider rule.
• Missing access or performance evidence remains unavailable.
• Findings distinguish eligibility from actual indexing, visibility, ranking, traffic, and value.
• Every remediation has an owner and a repeatable verification method.
• No ranking, indexing, rich result, traffic, or conversion outcome is guaranteed.

Handoff Signals

• Entity or claim ambiguity blocks interpretation → knowledge-entity-representation.
• Repeated generative retrieval or citation behavior must be studied → ai-discovery-measurement and generative-visibility-optimization.
• A finding requires code, infrastructure, content, redirect, metadata, or deployment changes → the relevant implementation owner.
• Access requires credentials, production logs, webmaster data, or policy approval → authorized consumer owner.
• A business prioritization decision remains unresolved → human or governance review.

Pairs Well With

• knowledge-entity-representation
• ai-discovery-measurement
• generative-visibility-optimization
• documentation
• architecture-review
• qa-review
• observability-review

Anti-patterns

• Treating a page response or manual browser view as proof of crawler or index state.
• Reporting eligibility, indexing, ranking, traffic, and conversion as the same outcome.
• Recommending every possible SEO tactic instead of the smallest evidence-backed remediation.
• Treating schema, a sitemap, or llms.txt as universally required.
• Ignoring authentication, environment, locale, duplication, rendering, or canonical conflicts.
• Inferring a performance problem without authorized search-performance evidence.
• Making machine access less safe, accessible, or understandable for people.