Claude SEO - Setup Guide¶
Source: skills.sh (31 skills, ~160,000 combined installs) GitHub: AgriciDaniel/claude-seo (16,628⭐, 2,436 forks, MIT, very active) Category: Marketing / SEO First Seen: Jul 23, 2026 sweep (rejected as "Claude Code-specific"); overturned Sep 9, 2026 - 31 skills are now on skills.sh in open Agent Skills format with broad cross-agent adoption Quality Tier: 🟡 Production (Gen Agent Trust Hub Pass / Socket Pass / Snyk Warn - W011 third-party content exposure, inherent to SEO workflows)
A universal SEO suite: 25 core skills + 6 extension skills, with the flagship seo-audit crawling up to 500 pages and delegating to up to 15 parallel specialist sub-agents. Covers technical SEO, E-E-A-T content quality, Schema.org, AI-search optimization (GEO/AEO), backlinks, local SEO, e-commerce, international SEO, and Google APIs - every audit output is a prioritized action plan with falsifiable recommendations grounded in Google primary-source guidance.
Installation¶
The skills themselves are open Agent Skills format and install via skills.sh:
npx skills add agricidaniel/claude-seo
For Hermes Agent, copy from a clone:
git clone --depth 1 https://github.com/AgriciDaniel/claude-seo.git
cp -r claude-seo/skills/* ~/.hermes/skills/
Heavy-tool dependency: several skills (e.g. seo-audit) call claude-seo run render_page.py - a CLI installed by the repo's setup script (isolated Python env + Playwright Chromium). Install it once per machine:
bash claude-seo/install.sh # from the clone; review before running
claude-seo doctor # verify the toolchain
(Claude Code users can instead use the plugin path: /plugin marketplace add AgriciDaniel/claude-seo → /plugin install claude-seo@agricidaniel-claude-seo → /seo setup. That path is Claude-Code-specific; the skills.sh + manual paths above are not.)
Prerequisites¶
| Requirement | Details |
|---|---|
| Hermes Agent | Skills directory support (~/.hermes/skills/) |
| Python 3 + Playwright | Installed by install.sh for the render tooling (headless Chromium for page rendering/SPA capture) |
| Node.js 18+ | For the npx skills path |
| Optional API keys | Extension skills: DataForSEO, Firecrawl, Ahrefs, Bing Webmaster, SE Ranking, Unlighthouse, Profound, Banana (image gen) - each extension is optional |
Skill Roster (core)¶
| Skill | Installs | What it does |
|---|---|---|
seo |
6,660 | Orchestrator - routes requests to the right sub-skill |
seo-audit |
6,027 | Full-site audit: crawls ≤500 pages, detects business type, delegates to ≤15 specialists, health score |
seo-content |
5,797 | E-E-A-T content quality analysis + optimization |
seo-competitor-pages |
5,747 | Competitor page-level analysis |
seo-sitemap / seo-technical |
5,613 / 5,427 | Sitemap audits / technical crawl diagnostics |
seo-geo |
5,399 | AI-search optimization: citability scoring, llms.txt, primary-source alignment |
seo-schema |
5,358 | Schema.org detect, validate, generate |
seo-page |
5,273 | Deep single-page analysis |
seo-backlinks / seo-plan |
5,247 / 5,232 | Backlink strategy / prioritized action planning |
seo-images / seo-hreflang / seo-programmatic / seo-local / seo-cluster / seo-google / seo-maps / seo-ecommerce / seo-drift / seo-sxo / seo-flow |
4.1K - 5.2K | Images, international, programmatic, local, semantic clustering, Google APIs, maps intelligence, e-commerce, ranking drift, search experience optimization, workflow engine |
seo-content-brief |
4,115 | Content briefs grounded in SERP evidence |
| Extensions | 58 - 69 | seo-firecrawl, seo-ahrefs, seo-dataforseo, seo-bing, seo-seranking, seo-profound, seo-unlighthouse, seo-image-gen |
Quick Start¶
- Install skills + run
install.shonce - "Audit corpusiq.io" →
seo-auditcrawls and produces a prioritized action plan - "Check our AI-search visibility" →
seo-geofor citability/llms.txt/AEO - "Write a content brief for [keyword]" →
seo-content-brief
CorpusIQ Use Cases¶
| Use Case | How |
|---|---|
| Docs SEO/AEO/GEO program | CorpusIQ's docs SEO pass (frontmatter, llms.txt, JSON-LD) gets automated audit + fix loops via seo-audit, seo-geo, seo-schema |
| Competitive SEO intelligence | seo-competitor-pages + seo-drift for tracking competitor ranking moves - feeds the competitive-attack playbook |
| Content briefs | seo-content-brief for SERP-grounded briefs before any blog/docs page is written |
| Technical SEO hygiene | seo-technical + seo-sitemap as a recurring site-health check for corpusiq.io and client domains |
Integration with CorpusIQ¶
CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.
This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.
Limitations / Verification¶
- Snyk Warn (W011, MEDIUM): skills ingest arbitrary third-party URLs (audited pages, SERPs) - indirect prompt-injection surface. Standard mitigation: treat page content as data, never as instructions; the skill's own workflow already separates analysis from actions
- Distribution is Claude-Code-first upstream (plugin marketplace,
/seoslash commands) - but the SKILL.md layer is platform-agnostic and that is what skills.sh installs - Extension skills are low-install on skills.sh (58 - 69) but functional; they just aren't widely adopted yet
# Verify skill installed
ls ~/.hermes/skills/seo-audit/SKILL.md ~/.hermes/skills/seo-geo/SKILL.md
# Verify toolchain
claude-seo doctor
Security¶
Verified Sep 9, 2026 on seo, seo-audit, seo-content:
| Audit | Verdict |
|---|---|
| Gen Agent Trust Hub | Pass |
| Socket | Pass |
| Snyk | Warn (MEDIUM W011 - third-party content exposure, inherent to SEO workflows) |
Related¶
← Skills Catalog | Marketplace → Powered by CorpusIQ