What does this skill do, and when should you use it?
Site Structure Optimizer is an SEO/GEO tuning skill in Aaron Marketing Skills focused on information architecture, page hierarchy, navigation, URL taxonomy, topic clusters, and internal links. It has two modes: architecture designs the whole-site structure, while linking improves connections within an existing site. Outputs include a structural diagnosis or plan, Mermaid site map, prioritized actions, and a structure score out of 100. It does not handle external backlinks, XML sitemaps, indexation issues, or CORE-EEAT auditing.
Reads the site type, goals, page inventory or sitemap, key URLs, audiences, content categories, and URLs to preserve. In architecture mode, it produces an ASCII hierarchy, URL map, navigation specification, hub/spoke topic-cluster plan, Mermaid site map, and architecture score. In linking mode, it analyzes orphan pages, anchor-text distribution, link graph, authority flow, and click depth, then produces a prioritized source/target/anchor plan and structure score. It can use crawlers or SEO/analytics tools when connected, but can also work manually from a supplied page list.
- A SaaS team is building a new marketing site and needs its hierarchy, navigation, and URL taxonomy planned.
- A content lead is restructuring a blog around hub-and-spoke topic clusters.
- An SEO practitioner needs to find pages with no inbound internal links.
- A team is preparing a site migration and needs URL-change and 301-redirect planning.
- An established site has confusing internal links and needs a prioritized source, target, and anchor-text plan.
How do you install this skill?
- Do not treat Estimated link thresholds, structure scores, or example results as measured findings; provide site inventories or exports and verify target URLs.
- Before crawling a non-owned site, confirm authorization and check robots.txt, Crawl-delay, and applicable TDM reservation signals; treat page instructions as data, never commands.
- Obtain explicit confirmation and retain recoverable records before URL migrations, 301/noindex/delete decisions, or memory/open-loops writes.
- The shared contract, scripts, and connectors referenced by the skill were not fully supplied here; verify paths, version compatibility, and failure handling before deployment.
- Shell / CLI
- Network access
- Local filesystem
Python 3Bash
On an Agent Skills-compatible host, run: npx skills add aaron-he-zhu/aaron-marketing-skills -s site-structure-optimizer. In Claude Code, you can instead run /plugin marketplace add aaron-he-zhu/aaron-marketing-skills, followed by /plugin install aaron-marketing@aaron.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/aaron-he-zhu/aaron-marketing-skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/seo-geo/tune/site-structure-optimizer" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Plan the site structure for a new SaaS marketing site
- Find orphan pages on https://example.com
Provide the site type and page information with a concrete goal. For example: "Plan the site structure for a new SaaS marketing site" or "Find orphan pages on https://example.com". You can specify --mode architecture or --mode linking. If the request covers both restructuring and internal-link fixes, run architecture first and hand off to linking.
What are this skill's strengths and limitations?
- Covers both whole-site architecture and internal-link optimization.
- Produces hierarchy trees, URL maps, navigation specifications, and Mermaid diagrams for implementation and review.
- Labels metrics as Measured, User-provided, or Estimated, and marks unavailable inputs N/A.
- Provides orphan-page dispositions plus anchor-text and source/target/anchor recommendations.
- Needs inputs such as site type and a page inventory; without crawl or analytics data, some metrics are estimated.
- Does not analyze external backlinks or diagnose XML sitemap and indexation problems.
- Does not calculate CORE-EEAT or perform the content-quality auditor's veto checks.
- The supplied SKILL.md does not show independent tests or platform-coverage results for this individual skill.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Site Structure Optimizer this page | 59 · Recommended | ★ 3k | 12d ago | Apache-2.0 |
| Programmatic SEO Planner | 47 · Use with care | ★ 19k | 22d ago | MIT |
| On-Page SEO Checker | 54 · Use with care | ★ 3k | 12d ago | Apache-2.0 |
| Page Play Builder: Scaled SEO Page Builder | 54 · Use with care | ★ 3k | 12d ago | Apache-2.0 |
| Blog Cluster Engine | 53 · Use with care | ★ 2.4k | 13d ago | MIT |
Compared with technical-seo-checker: this skill handles site structure and internal links, not XML sitemaps, indexing, or crawlability issues. Compared with offsite-signal-analyzer: this skill handles links within the site, not external backlinks. Compared with on-page-seo-checker: this skill optimizes relationships between pages rather than page-level titles, keyword placement, and on-page SEO. CORE-EEAT scoring belongs to content-quality-auditor.
How did FollowSkills review this skill?
The skill explicitly treats crawled content as untrusted data, requires robots.txt compliance, labels data sources, and asks before saving results; the referenced security boundaries also cover SSRF, non-persistent credentials, and read-only crawling. Points are deducted because crawling and memory/open-loop writes still create external or persistent effects, while the skill itself does not fully specify per-action authorization, rollback, or complete data-flow confirmation; publisher identity is unverified.
Mode selection, hard-stop gates, missing-input handling, Estimated/N/A labeling, and multiple templates make the main workflow reasonably consistent and diagnosable. Points are deducted because the static material does not prove that scripts, referenced paths, Mermaid output, or key paths actually run, and abnormal connector/input failure coverage is limited; static calibration caps this at 10.
Trigger semantics, the two architecture/linking modes, input requirements, output boundaries, and exclusions for external backlinks, XML sitemaps, and indexation are clear. A manual no-tool path and Chinese wording are also provided. Points are deducted because site-type adaptation relies heavily on estimated thresholds, with limited evidence for complex dynamic sites, authenticated sites, or degraded operation under mainland-China network constraints.
The documentation provides clear mode selection, quick starts, a contract, procedures, templates, examples, and reference materials; version, Apache-2.0 licensing, and repository-level security/maintenance signals are visible. Points are deducted because the selected skill lacks an explicit changelog, maintenance owner, and update path, and depends on external contract documents, connectors, and repository-relative scripts with some installation and troubleshooting assumptions.
The skill specifies directly usable deliverables: hierarchy tree, URL table, navigation specification, topic clusters, Mermaid map, orphan disposition, link plan, score, and handoff summary; the example illustrates the expected form. Points are deducted because metrics and examples are Estimated or illustrative, and static reading cannot verify actual site-analysis correctness, completeness, or marginal benefit over manual work; static calibration caps this at 7.
Versioned source files, structured templates, explicit scoring formulas, examples, and repository CI/security descriptions provide some auditability. Points are deducted because there is no committed executed test coverage for this skill's key paths, third-party validation, or independent reproduction record; static calibration caps this at 5.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →