SEO Architecture

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. Include concise robots/XML-sitemap discovery comments and llms.txt navigation sentences pointing to actual Markdown answers; preserve the native protocols and verify delivery as described in the local contract.

Lane outputs: site-architecture.md; navigation-map.md (combine small related steps with stable section links). Build llms.txt -> sitemap.md -> topic map -> substantive page.md paths with descriptive direct destinations. Audit orphans, cycles, collisions, dead links and question-to-answer task completion. Avoid forced traversal through irrelevant HTML navigation.

Create and audit a bounded site-graph.json that explains how users and permitted crawlers discover important pages, how templates connect, and where query/entity coverage lives.

Read site-graph-protocol.md before a formal architecture audit.

Procedure

1. Define domain/subdomain scope, environments, URL patterns, crawl authorization, seed pages, locale, user tasks, business-critical templates, exclusions, and collection limits.
2. Capture nodes and crawlable link edges with source/target URLs, link location, anchor or accessible name, follow state, template/context classification, and observation time. A supported crawler CSV can be normalized with python "<suite-root>/scripts/import_seo_exports.py" crawler-csv ... using <suite-root>/docs/DATA-IMPORT-ADAPTERS.md, but that observation envelope is not a Site Graph: bind every node and edge to the raw captures and metadata required by the Site Graph contract.
3. 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. Validate the graph with python "<suite-root>/scripts/validate_site_graph.py" validate-graph <bundle>/site-graph.json --bundle <bundle>.
4. Inspect disconnected/orphan nodes, inbound-path evidence, hubs, crawl paths, pagination, facets, filters, parameter spaces, taxonomy, duplicate routes, canonical/index conflicts, and locale clusters.
5. Map the formal Query Corpus and entity/page ownership to nodes. Do not create a page or link merely because a keyword exists.
6. Propose the smallest user-meaningful navigation/contextual-link change. Avoid sitewide boilerplate links, manipulative anchors, hidden links, or arbitrary depth/count thresholds.
7. Route crawl/index defects to seo-technical, page/content ownership to seo-page/seo-content, programmatic templates to seo-programmatic, and authority work to seo-authority.
8. Package findings as SEO Findings 1.1.0 with site_graph, site, template, or web_page targets and architecture, technical, content, or international categories.

Formal evidence handoff

After graph validation, package only supported observations as SEO Findings 1.1.0. Using the same <suite-root> definition above, run python "<suite-root>/scripts/validate_seo_findings.py" validate-findings <bundle>/seo-findings.json --bundle <bundle>. Keep the handoff provisional until both validators pass.

Output

Return the validated graph, scope and coverage, important node/edge findings, orphan/discovery evidence, taxonomy/facet risks, query/entity mapping, owners, verification, rollback, and limitations. Do not publish a composite authority or architecture score.