SEO Audit

SEO Audit is one individual agent skill from Corey Haines' open-source Marketing Skills collection. It does not give an AI agent access to a Search Console property, and it does not replace a crawler. Instead, it supplies a clear review workflow an agent can use to assess a website, available data, and access supplied by the team. According to the repository, the collection is for technical marketers and founders using agents for conversion optimization, copywriting, SEO, analytics, and growth engineering. This is the appropriate skill when the task is not writing one page but documenting, evidencing, and prioritizing technical or content SEO problems.

What the audit workflow covers

The skill organizes an investigation around technical SEO, on-page factors, and content. Its technical checks include crawlability and indexability, robots.txt, XML sitemaps, status codes, redirects, canonicals, JavaScript rendering, performance, mobile usability, and structured data. For international or multilingual sites, the guidance also calls out hreflang, return tags, self-references, x-default, canonical URLs, and localized content as common failure points. That makes it useful when a team wants to turn weak organic visibility into a testable list of possible causes rather than merely watch rankings.

The on-page section examines titles, meta descriptions, heading hierarchy, search intent, internal links, images, and alt text. The content review considers freshness, topical coverage, intent fit, cannibalization, authorship, and trust signals. The skill distinguishes SaaS, ecommerce, blog, local-business, and international sites. That distinction helps avoid treating a faceted navigation like an editorial publication or a missing location page like a product-detail issue.

An output, not merely a checklist

The guidance asks for an executive summary, concrete findings with impact, evidence, fix, and priority, plus an action plan sorted by urgency. This is useful when handing work to engineering, content, or product marketing: “check the canonical” becomes actionable only when the URL, observed behavior, SEO consequence, and next step are clear. An agent should not invent evidence, however. Without crawl data, Search Console, server logs, or rendered pages, a conclusion remains a hypothesis and should be labelled accordingly.

Installation, data, and boundaries

The Marketing Skills library is free and MIT-licensed according to its repository. The direct Claude Code command for this individual skill is:

npx skills add coreyhaines31/marketingskills --skill seo-audit -a claude-code

The command installs guidance, not SEO tooling or permissions automatically. Useful inputs include a target URL, a crawl export, Search Console data, analytics metrics, and context about important pages or migrations. A browser or rendering tool matters because a plain HTML fetch can miss JavaScript-injected JSON-LD. Costs for optional third-party tools or provider plans are outside this entry; consult those providers for current terms.

Scope and freshness

The official product site is marketing-skills.com and the source is github.com/coreyhaines31/marketingskills. The repository had 47,560 GitHub stars on 2026-09-07. That count belongs to the collection, not SEO Audit alone, and is only a point-in-time popularity signal—not evidence of quality, security, or SEO results. Before relying on an audit in production, a team should validate generated findings against real URLs, measurement data, and the applicable search-engine guidance.