Claude Code Skills for SEO: The Complete Guide (2026)

Short answer: install AgriciDaniel’s free claude-seo plugin for audits, technical SEO, schema and GEO, add Corey Haines’ free marketingskills for content and on-page work, and connect one data source (Google Search Console through mcp-gsc, or DataForSEO) so the skills work from real numbers instead of guesses. That stack covers most SEO jobs at no cost beyond your Claude plan and any data API.

This is the broad guide: what SEO skills are, which ones fit each SEO job (audits, content, keyword research, schema, GEO, programmatic), the MCP data servers they rely on, install steps, stacks by role and the mistakes to avoid. If you only want a scored ranking, read the best Claude Code skills for SEO, ranked. If you are running an audit today, go to Claude Code skills for SEO audits.

A skill is a folder of instructions, scripts and reference files that Claude loads when the task matches. It is only as good as the data it can reach: a keyword research skill with no keyword data is guessing. Every repo, command and price below was checked against its source page on 1 November 2026, with GitHub stars and update dates from the GitHub API on that date. Where a price was not public, it says so.

What are Claude Code skills for SEO?

Claude Code skills for SEO are SKILL.md folders that teach Claude a specific SEO workflow, such as a technical audit, a schema build or a keyword cluster plan. They live in ~/.claude/skills (personal) or .claude/skills (project), or arrive bundled inside plugins. Claude reads each skill’s description and loads the full instructions only when your request matches.

A skill folder usually holds:

SKILL.md: a name, a description that tells Claude when to use it, and step-by-step instructions
references/: checklists, rubrics, schema templates, Google guideline summaries
scripts/: Python or Node helpers that fetch pages, parse HTML or call APIs

That last part matters for SEO. A skill can ship a script that actually fetches your sitemap and counts status codes, so Claude reports what it found instead of what it imagines.

If you are new to the format, start with what Claude Code skills are and how they load. If you have never touched a terminal, Claude Code skills for non-developers covers the setup without assuming you code.

Skills vs prompts vs MCP servers

Layer	What it gives Claude	SEO example
Prompt	A one-off instruction	”Audit this page”
Skill	A reusable method, rubric and scripts	A 50-check technical audit with a scoring rubric
MCP server	Live data and actions from an outside tool	Search Console clicks, DataForSEO search volume
Plugin	A bundle of skills, commands, agents and MCP config	claude-seo, Anthropic’s Marketing plugin

Takeaway: skills supply the method, MCP servers supply the data. You usually want both.

How we picked these skills

There are thousands of SEO skills scattered across directories like claudemarketplaces.com, claudskills.com and skills.sh. Many are a single SKILL.md with a paragraph of generic advice. We filtered with five questions:

Is the source public and inspectable? You should be able to read every file before it runs on your machine.
Is it maintained? We noted the last push date from GitHub. A skill untouched for a year may describe a Google that no longer exists.
Does it do something a prompt cannot? Scripts, rubrics, reference files or data connections count. A rewritten prompt does not.
Does it cover a real SEO job? Audits, keyword research, technical fixes, schema, content, GEO or programmatic SEO.
Is the install path clear? One command, or a short documented sequence.

Star counts are a popularity signal, not a quality score. A repo with 200 stars and tight scripts can beat one with 50,000 stars and thin instructions. We list stars so you can see adoption, then describe what the skill actually contains.

What we did not do: we have not published timed benchmarks or before-and-after ranking data for these tools. Anyone selling you “rank #1 in 30 days with this skill” is making a claim no skill can back.

The 14 Claude Code skills for SEO covered in this guide

#	Skill or source	Free or paid	Best for	GitHub stars (1 Nov 2026)	Last update
1	AgriciDaniel/claude-seo	Free (MIT)	Full-site audits, technical, schema, GEO	25,377	1 Nov 2026
2	coreyhaines31/marketingskills	Free (MIT)	SEO audit, AI SEO, programmatic SEO, schema	57,802	31 Oct 2026
3	TheCraigHewitt/seomachine	Free (MIT)	Research-to-publish blog workflow with GSC, GA4, DataForSEO	7,565	23 Oct 2026
4	AgriciDaniel/claude-blog	Free (MIT)	Blog writing, briefs, decay, cannibalization	2,606	30 Oct 2026
5	alirezarezvani/claude-skills	Free (MIT)	AEO and local SEO inside a large marketing bundle	29,110	31 Oct 2026
6	Anthropic Marketing plugin	Free plugin (Apache-2.0)	Content, competitor briefs, Ahrefs and SimilarWeb connectors	34,324 (repo)	1 Nov 2026
7	aaron-he-zhu/seo-geo-claude-skills	Free (Apache-2.0)	SEO plus GEO skill set	233	4 Sep 2026
8	ComposioHQ/awesome-claude-skills	Free directory	Finding more skills	78,500	18 Sep 2026
9	DataForSEO MCP server	Free server, paid API	Keyword, SERP, backlink data	258	1 Nov 2026
10	AminForou/mcp-gsc	Free	Google Search Console data	2,086	15 Sep 2026
11	Ahrefs MCP	Paid (Ahrefs plan)	Backlinks, keywords, competitor data	n/a (hosted)	n/a
12	Semrush MCP	Paid (Semrush plan)	Keyword and domain analytics	n/a (hosted)	n/a
13	Ultimate Claude SEO and AEO Skills Pack (Contra)	Paid, $39 listed	Packaged SEO and AEO skill files	n/a	n/a
14	Master Marketing Skills Bundle (Gumroad)	Paid, $167 one-time	Large marketing bundle with SEO section	n/a	n/a

The next sections go through each one: what it does, how to install it, and where it falls short.

Best all-in-one SEO skill suites

If you want one install that covers most of SEO, start here. These are the suites with the broadest coverage and the most active maintenance.

1. claude-seo by AgriciDaniel

Source: github.com/AgriciDaniel/claude-seo · Price: free, MIT license · Stars: 25,377

claude-seo is the most complete open-source SEO suite for Claude Code we found. Its README describes 26 sub-skills and 19 sub-agents behind a single /seo command family.

What it covers:

Audits and planning: /seo audit <url>, /seo page <url>, /seo plan <type>, plus /seo setup and /seo doctor for configuration checks
Technical SEO: /seo technical, /seo sitemap, /seo images, /seo hreflang, /seo programmatic, /seo drift
Content and E-E-A-T: /seo content, /seo content-brief, /seo cluster, /seo sxo, /seo competitor-pages
Schema and AI search: /seo schema, /seo geo, /seo agentic
Local, links and commerce: /seo local, /seo maps, /seo backlinks, /seo ecommerce
Extensions: SE Ranking, Profound, Bing Webmaster, Unlighthouse and an image generator, each with its own setup guide

Install inside Claude Code (version 1.0.33 or newer):

/plugin marketplace add AgriciDaniel/claude-seo
/plugin install claude-seo@agricidaniel-claude-seo
/seo setup

A manual install via git clone and install.sh (or install.ps1 on Windows) is also documented.

Where it falls short: the breadth is the risk. Thirty commands is a lot to learn, and the parallel sub-agents can use a large share of your Claude usage on a big audit. The README also notes a private paid version for members of the author’s community, so some features may land there first.

Best for: SEO leads and agencies who want one toolkit for audits, technical checks, schema and GEO.

2. marketingskills by Corey Haines

Source: github.com/coreyhaines31/marketingskills · Price: free, MIT · Stars: 57,802

This is a marketing library, not an SEO-only suite, but its “SEO and Discovery” group is one of the cleanest starting points. It includes:

seo-audit
ai-seo (optimizing for AI answers and citations)
programmatic-seo
schema
site-architecture
link-building
competitors

Install everything, or just the skills you want:

npx skills add coreyhaines31/marketingskills
npx skills add coreyhaines31/marketingskills --skill seo-audit programmatic-seo schema

Where it falls short: these skills are method-heavy and data-light. They tell Claude how to think about an audit or a programmatic page set, but they expect you to supply the data or connect an MCP server.

Best for: marketers who want SEO skills that sit alongside copywriting, CRO and launch skills. Our guide to Claude Code skills for marketers covers the non-SEO half of this library.

3. claude-skills by Alireza Rezvani

Source: github.com/alirezarezvani/claude-skills · Price: free, MIT · Stars: 29,110

A very large multi-domain library. The marketing domain lists 50 skills across 8 groups. The two clearly SEO-focused ones are aeo (answer engine optimization, with an E-E-A-T audit and citation tracking across several LLMs) and local-seo-manager (Google Business Profile, NAP consistency, Map Pack). An seo-audit skill appears in its orchestration examples.

Install the marketing bundle in Claude Code:

/plugin marketplace add alirezarezvani/claude-skills
/plugin install marketing-skills@claude-code-skills

Where it falls short: SEO is a small slice of a huge repo. You will install many skills you do not need just to get two or three.

Best for: local businesses and teams that care about AI answer visibility.

Content is where most SEO hours go, and where generic AI output does the most damage. These two projects build structure around the writing so it is not a single “write me a blog post” prompt.

4. SEO Machine by Craig Hewitt

Source: github.com/TheCraigHewitt/seomachine · Price: free, MIT (data sources cost extra) · Stars: 7,565

SEO Machine is a full Claude Code workspace rather than a single skill. It was originally built for the podcast host Castos and is now open source. You clone the repo and run Claude inside it.

Commands include:

Content: /research, /write, /rewrite, /analyze-existing, /optimize, /article, /priorities, /publish-draft, and /scrub to remove AI-style patterns
Research: /research-serp, /research-gaps, /research-trending, /research-performance, /research-topics
Landing pages: /landing-write, /landing-audit, /landing-research, /landing-competitor

The standout is its data layer. It connects to Google Analytics 4, Google Search Console and DataForSEO, so /research-performance can work from your real clicks and rankings.

Install:

git clone https://github.com/TheCraigHewitt/seomachine.git
cd seomachine
pip install -r data_sources/requirements.txt
claude

Then fill in the context/ templates (brand voice, style guide, target keywords, internal links). The examples/castos/ folder shows a finished set.

Where it falls short: setup takes real time. You need credentials for GA4, GSC and a DataForSEO account, and the context files only help if you fill them in properly.

Best for: content teams publishing several posts a week who want research, drafting and optimization in one place. See Claude Code skills for content writers for writing-first alternatives.

5. claude-blog by AgriciDaniel

Source: github.com/AgriciDaniel/claude-blog · Price: free, MIT · Stars: 2,606

From the same author as claude-seo, this suite covers the blog lifecycle with 30 /blog subcommands:

Writing: write, rewrite, outline, brief
Planning: strategy, calendar, cluster
Quality and SEO: analyze (0 to 100 score), seo-check, geo, audit, cannibalization, factcheck, decay
Publishing: schema, taxonomy, repurpose, plus translation and localization commands
Integrations: google for PageSpeed, Search Console and GA4

Each draft passes a five-gate check scored against a 100-point rubric, and delivery is blocked below 90.

/plugin marketplace add AgriciDaniel/claude-blog
/plugin install claude-blog@agricidaniel-blog

The README recommends pinning a reviewed commit or release tag rather than tracking main. That is good advice for any skill that runs scripts.

Where it falls short: a rubric score is the skill grading its own work. Treat a 92/100 as “passed the checklist”, not “will rank”.

Best for: solo bloggers and small teams who want briefs, writing, decay checks and cannibalization in one tool.

Best skills for keyword research

Keyword research is the job where skills without data fail hardest. Claude has no live search volume. Any number it produces from memory is a guess.

So the honest answer is: the best keyword research setup is a method skill plus a data MCP server. For the method, use the competitors and site-architecture skills from marketingskills, /seo cluster from claude-seo, or /research-topics and /research-gaps from SEO Machine. For the data, connect DataForSEO, Ahrefs or Semrush (covered below).

A workable keyword flow in Claude Code looks like this:

Seed: give Claude your product, three competitors and your Search Console export.
Expand: the MCP server pulls related keywords, volumes and difficulty.
Classify: the skill sorts terms by intent (informational, commercial, transactional) and by funnel stage.
Cluster: group terms that one page can rank for together, based on overlapping SERP results.
Prioritize: score clusters by business fit, difficulty and existing rankings.
Map: assign each cluster to an existing URL or a new page.

Step	Needs live data?	Good source
Seed list	No	Your own knowledge, GSC export
Volumes and difficulty	Yes	DataForSEO, Ahrefs, Semrush
Intent classification	Partly	SERP data plus skill rubric
Clustering by SERP overlap	Yes	DataForSEO SERP endpoints
Prioritization	No	Skill rubric plus your judgment

We go deeper on prompts, clustering logic and cost control in Claude Code skills for keyword research.

Takeaway: if a keyword skill gives you search volumes without a connected data source, those numbers are invented.

Best skills for SEO audits and technical SEO

Audits are where Claude Code skills shine, because a good audit is a checklist plus judgment, and skills are built for checklists.

What a strong audit skill checks

Area	Example checks
Crawlability	robots.txt rules, XML sitemap validity, blocked resources
Indexability	noindex tags, canonical conflicts, redirect chains
On-page	title and meta length, H1 presence, duplicate titles
Performance	Core Web Vitals (LCP, INP, CLS) via PageSpeed or Lighthouse
Internationalization	hreflang return tags, language codes
Structured data	valid JSON-LD, required properties, type fit
Content	thin pages, duplicate content, E-E-A-T signals
AI crawlers	robots rules for GPTBot, ClaudeBot, PerplexityBot and others

Our top picks:

claude-seo for depth: /seo audit, /seo technical, /seo sitemap, /seo hreflang, /seo images, plus the Unlighthouse extension for Lighthouse scores across every page.
marketingskills seo-audit for a lighter, method-first audit you can run on a single page or a small site.
claude-blog audit and cannibalization for content-side audits on a blog.

Where audit skills go wrong

A skill that cannot fetch your pages will audit from assumptions. Before you trust any audit, check that the output cites specific URLs and specific findings (“/pricing has two canonical tags”) rather than general advice (“ensure canonical tags are correct”).

Large sites are a second problem. Claude Code is not a crawler. For sites with tens of thousands of URLs, export a crawl from Screaming Frog or a similar tool and hand the skill the CSV. The skill then does what it is good at: reading, grouping and prioritizing.

The full walkthrough, including a sample audit prompt and how to read the output, is in Claude Code skills for SEO audits.

Best skills for schema markup

Schema is the most mechanical SEO task, which makes it a great fit for skills. A schema skill holds templates for each type, the required and recommended properties, and rules for which types fit which pages.

Options:

claude-seo /seo schema: detects, validates and generates JSON-LD.
marketingskills schema: a focused schema skill you can install on its own with --skill schema.
claude-blog schema: generates Article, FAQ and related markup for posts.

A good schema workflow:

Ask the skill to detect existing markup on the page.
Have it propose the right types (Article, Product, FAQPage, LocalBusiness, Organization, BreadcrumbList).
Generate JSON-LD using only facts that appear on the page.
Validate with Google’s Rich Results Test and the Schema.org validator.

That third step is the one people skip. Schema that claims reviews, prices or ratings not shown on the page breaks Google’s structured data guidelines. A schema skill makes it easy to generate markup fast, so it also makes it easy to generate wrong markup fast.

Takeaway: let the skill write the JSON-LD, then validate it yourself every time.

Best skills for GEO and AI search visibility

GEO (generative engine optimization) is the work of getting your pages cited in AI answers from ChatGPT, Perplexity, Gemini, Claude and Google’s AI Overviews. It overlaps heavily with good SEO: clear answers, named entities, sources, structured pages.

Skills that target it:

claude-seo /seo geo and /seo agentic, plus the SE Ranking and Profound extensions, which track AI share of voice and AI citations.
marketingskills ai-seo, a method skill for writing and restructuring pages so AI engines can extract and cite them.
alirezarezvani aeo, with an E-E-A-T audit and citation tracking across several LLMs.
aaron-he-zhu/seo-geo-claude-skills (GitHub, free, Apache-2.0, 233 stars, last push 4 September 2026), a smaller SEO plus GEO skill set.
claude-blog geo, which scores a post’s readiness for AI citation.

What GEO skills actually change on a page

Change	Why it helps AI extraction
A 40 to 80 word direct answer under each H2	Easy for an engine to lift as a cited passage
Named entities and sources	Lets the engine verify and attribute
Tables for comparisons	Structured data the model can read cleanly
FAQ blocks from real questions	Matches how people phrase AI prompts
AI crawler access in robots.txt	The page cannot be cited if it cannot be fetched
An llms.txt file	A proposed standard; adoption by AI engines is still unclear

Be careful with claims here. AI citation is hard to measure, and tracking tools sample a limited set of prompts. A GEO skill can make your pages more extractable. It cannot promise that a given engine will cite you.

Best skills for programmatic SEO

Programmatic SEO means generating many pages from a template and a dataset: “[tool] alternatives”, “[city] plumbers”, “[integration A] + [integration B]”. It works when every page answers a real query with real, distinct data. It fails, and can trigger Google’s scaled content abuse policy, when the pages are thin variations.

Skills to use:

marketingskills programmatic-seo: covers choosing a pattern, building the dataset, designing the template and avoiding thin pages.
claude-seo /seo programmatic: analyzes and plans programmatic page sets, with quality gates.
marketingskills site-architecture: for planning how the generated pages link together.

A safe programmatic process:

Prove demand: confirm the pattern has search volume across many variants (DataForSEO or Ahrefs).
Own the data: each page needs at least one data point competitors do not have.
Build one page by hand first: if it is not useful as a single page, it will not be useful as 500.
Template with variation: sections should change based on the data, not just swap the keyword.
Launch in batches: publish 20 to 50, watch indexing in Search Console, then scale.

Takeaway: the skill helps you plan and template. The unique data is your job, and it is the part that decides whether the pages survive.

The MCP data sources these SEO skills rely on

A skill without data reasons from what Claude already knows. A skill with an MCP server reasons from your real numbers. These are the servers that matter for SEO.

9. DataForSEO MCP server

Source: github.com/dataforseo/mcp-server-typescript · Price: the server is free (Apache-2.0); the DataForSEO API is paid per use, see dataforseo.com/pricing · Stars: 258

DataForSEO is the cheapest route to raw keyword, SERP, backlink and on-page data for most small teams, because you pay per request rather than a monthly seat. The current server (v3) gives Claude four tools: three for reading DataForSEO’s documentation and one, api_request, for making authenticated calls. That design means Claude looks up the right endpoint, then calls it.

Two ways to connect:

{
  "mcpServers": {
    "dataforseo": { "url": "https://mcp.dataforseo.com/v3/mcp" }
  }
}

The remote URL uses OAuth. Or run it locally with npx dataforseo-mcp-server and set DATAFORSEO_LOGIN and DATAFORSEO_PASSWORD. Get credentials at app.dataforseo.com/api-access.

Watch the bill: an agent that loops over hundreds of keywords makes hundreds of requests. Set a spending limit in your DataForSEO account before you run a large job.

10. mcp-gsc for Google Search Console

Source: github.com/AminForou/mcp-gsc · Price: free · Stars: 2,086 · Last push: 15 September 2026

Search Console is your only first-party record of how Google sees your site: real queries, clicks, impressions, positions and index status. mcp-gsc connects it to Claude using OAuth or a service account. Once connected, you can ask for queries with high impressions and low CTR, pages that lost clicks month over month, or URLs with indexing issues.

Google does not publish an official Search Console MCP server, so every GSC connector is a community project. Read the code and scope the permissions before you grant access. Note that a different package called mcp-gsc exists on PyPI that the project says it does not maintain, so install from the GitHub instructions.

11. Ahrefs MCP

Source: docs.ahrefs.com/mcp · Price: paid; requires an Ahrefs subscription (Ahrefs documents remote MCP access on paid plans from Lite up; current plan prices at ahrefs.com/pricing)

Ahrefs runs a hosted remote MCP server that works with Claude Code. It exposes Ahrefs data such as backlinks, keywords and competitor metrics, and calls draw on your plan’s API units. Ahrefs also publishes a local server on GitHub that works with API v3 keys.

12. Semrush MCP

Source: Semrush’s official documentation · Price: paid; requires a qualifying Semrush plan, pricing on semrush.com/prices

Semrush offers an official remote MCP server with OAuth sign-in. Third-party guides disagree on the endpoint version, so copy the current URL from Semrush’s own docs. Like Ahrefs, it consumes API units from your subscription.

Which data source should you connect?

Data source	Cost model	Best for	Official?
DataForSEO MCP	Pay per API call	Keyword volumes, SERPs, backlinks on a budget	Yes, from DataForSEO
mcp-gsc	Free	Your own queries, clicks, index status	Community (no official Google server)
Ahrefs MCP	Existing Ahrefs plan	Backlink and competitor depth	Yes, from Ahrefs
Semrush MCP	Existing Semrush plan	Keyword and domain analytics	Yes, from Semrush
GA4 (via SEO Machine or claude-blog)	Free	Engagement and conversions	Built into those projects

If you already pay for Ahrefs or Semrush, use that. If you do not, DataForSEO plus Search Console covers most needs at the lowest starting cost.

Paid Claude Code SEO skills: are they worth it?

Paid SEO skill packs exist on Gumroad, Contra and standalone storefronts. We checked the listings we could verify:

13. The Ultimate Claude SEO and AEO skills pack on Contra

Source: contra.com listing · Price: $39 listed (marked down from $79) · Seller: Samuel Uebersax

Twelve skill files covering site audits, SEO and AEO writing, content briefs, JSON-LD schema, keyword intent classification, internal linking, competitor gaps, E-E-A-T scoring, AI visibility checks and site architecture, plus two Webflow-specific bonus skills and a setup guide. Install by copying the folders into ~/.claude/skills.

14. The Master Marketing skills bundle on Gumroad

Source: inflectual.gumroad.com · Price: $167 one-time per the listing we could read

Titled as 167 Claude skills for marketing. We could not confirm the price or the exact number of SEO skills from the page, so check the listing before buying.

Free vs paid: the honest comparison

Factor	Free (GitHub)	Paid packs
Price	$0 (data APIs extra)	Typically tens of dollars one-time, where listed
Inspect before install	Yes, full source	Usually only after purchase
Update history	Visible commits and dates	Promised, not visible
Scripts and data hooks	Common in the big suites	Often markdown-only
Community issues and fixes	Public issue trackers	Seller support only
Niche fit (e.g. Webflow)	Varies	Sometimes better
License clarity	MIT or Apache-2.0 on our picks	Check the terms

The free suites above are larger and more actively maintained than any paid pack we could inspect. A paid pack makes sense in two cases: it targets your exact platform (such as Webflow), or you value a curated, short set over a large library. Either way, read every file before running it.

How to install Claude Code skills for SEO

There are three install paths. Pick the one the repo documents.

Method	Command	Use when
skills CLI	`npx skills add owner/repo`	The repo is a plain skills collection (e.g. marketingskills)
Plugin marketplace	`/plugin marketplace add owner/repo` then `/plugin install name@marketplace`	The repo ships as a plugin (claude-seo, claude-blog, Anthropic plugins)
Manual	Copy the skill folder into `~/.claude/skills/` or `.claude/skills/`	Paid packs, single skills, or anything you want to audit first

Personal or project? Put skills you use everywhere in ~/.claude/skills. Put client-specific skills (with that client’s brand voice and keyword lists) in the project’s .claude/skills so they travel with the repo and stay out of other work.

Anthropic’s official Marketing plugin

Source: github.com/anthropics/knowledge-work-plugins · Price: free, Apache-2.0

Anthropic’s own Marketing plugin is aimed at content drafting, campaign planning, brand voice, competitor briefs and performance reporting. Its listed connectors include Ahrefs and SimilarWeb, which is where its SEO value comes from. It is not an SEO audit suite.

claude plugin marketplace add anthropics/knowledge-work-plugins
claude plugin install marketing@knowledge-work-plugins

For a step-by-step version with screenshots of each path, see how to install Claude Code skills.

A safety checklist before you install

Read SKILL.md and every script. Skills can run code on your machine.
Pin a release tag or commit when the repo offers one.
Give MCP servers the narrowest credentials possible. A read-only GSC user is enough.
Set API spending caps on DataForSEO, Ahrefs or Semrush before long agent runs.
Keep client data in project-scoped skills, not global ones.

Common mistakes when using Claude Code for SEO

These come up repeatedly in Reddit threads and community discussions about AI SEO workflows.

Trusting numbers with no data source. If no MCP server is connected, any search volume or difficulty score is made up.
Installing every skill. Overlapping skills with similar descriptions confuse Claude about which to load. Install for the job you have.
Skipping the context files. Suites like SEO Machine depend on your brand voice, keywords and internal link list. Empty templates produce generic content.
Publishing AI drafts unedited. Google’s guidance targets content made mainly to manipulate rankings, whatever made it. Unedited, unoriginal drafts at scale are the risk.
Treating self-scores as outcomes. A rubric score of 95 means the checklist passed. Rankings are decided by Google.
Auditing huge sites inside the chat. Use a crawler for collection and the skill for analysis.
Forgetting cost. Parallel sub-agents use more of your Claude plan, and paid APIs charge per call.
Never updating. SEO guidance changes. Pull updates on a schedule and re-read the changelog.

Build your own SEO skill when the list does not fit

Off-the-shelf skills encode someone else’s process. Your agency’s audit template, your client reporting format or your editorial checklist probably differ.

A custom SEO skill can be small:

~/.claude/skills/client-seo-report/
  SKILL.md          # when to use it, the steps, the output format
  references/
    checklist.md    # your audit checks
    report-template.md
  scripts/
    gsc_pull.py     # optional: pull last 28 days from Search Console

Write the description field carefully. It is how Claude decides to load the skill, so name the trigger phrases you actually type (“monthly SEO report”, “client audit”).

The full process, including testing and sharing, is in how to build a Claude Code skill. Copywriters building their own variants should also see Claude Code skills for copywriters.

Recommended SEO skill stacks by role

You are	Install	Connect
Solo blogger	claude-blog	mcp-gsc
In-house content team	SEO Machine	GSC, GA4, DataForSEO
Technical SEO	claude-seo	mcp-gsc, DataForSEO or Ahrefs
Agency	claude-seo plus a custom report skill	Your existing Ahrefs or Semrush plan, client GSC
Marketer who does some SEO	marketingskills (SEO group) plus Anthropic Marketing plugin	Ahrefs if you have it
Local business	alirezarezvani `local-seo-manager`	mcp-gsc
Programmatic SEO builder	marketingskills `programmatic-seo` plus claude-seo	DataForSEO

Start with one suite and one data source. Run it on a real task this week, such as a single page audit or one content brief. Add more only when you hit a gap.

FAQ

What are the best Claude Code skills for SEO?

For full coverage, claude-seo by AgriciDaniel is the broadest free suite. For marketers, the SEO group in coreyhaines31/marketingskills is a lighter option. For content teams, SEO Machine and claude-blog cover research to publishing.

Can Claude Code do SEO?

Yes, for analysis, planning and writing. It can audit pages, generate schema, cluster keywords and draft content. It needs MCP servers or exported files for live data such as search volume and rankings.

Can Claude do keyword research?

Only with a data source. Connect DataForSEO, Ahrefs or Semrush through MCP, or paste an export. Without one, any volume it reports is a guess.

Is there a Google Search Console MCP for Claude?

Yes, community ones. AminForou/mcp-gsc is a popular open-source option. Google does not publish an official Search Console MCP server.

Does DataForSEO have an MCP server?

Yes. DataForSEO maintains an official server on GitHub (dataforseo/mcp-server-typescript) with a remote URL and a local npx option. The API itself is billed per use.

Do Ahrefs and Semrush work with Claude Code?

Both offer official remote MCP servers that need a paid plan. Calls consume API units from your subscription.

Are Claude Code SEO skills free?

Most of the strongest ones are free and open source under MIT or Apache-2.0. Paid packs exist, such as a $39 SEO and AEO pack on Contra. Data APIs usually cost money.

How do I install an SEO skill in Claude Code?

Use npx skills add owner/repo for plain skill repos, /plugin marketplace add then /plugin install for plugins, or copy the folder into ~/.claude/skills.

Where are Claude Code skills stored?

Personal skills go in ~/.claude/skills. Project skills go in .claude/skills inside the repo. Plugins install their skills automatically.

What is the difference between a skill and an MCP server?

A skill is instructions and scripts that teach Claude a method. An MCP server connects Claude to an outside tool or dataset. SEO workflows usually need both.

Can Claude Code skills help with GEO and AI Overviews?

They can restructure pages so AI engines can extract and cite them: direct answers, tables, entities and sources. No skill can guarantee a citation from any engine.

Google’s stated focus is helpfulness, not how content is made. Mass-producing unoriginal pages to manipulate rankings breaks its spam policies regardless of the tool.

Can Claude Code replace Screaming Frog or Ahrefs?

No. It is not a crawler or a link index. It works best reading crawl exports and tool data, then grouping, prioritizing and writing fixes.

Do I need to know how to code to use these skills?

No, but you need to be comfortable typing commands into a terminal. Plugin installs are two commands. Setting up MCP credentials takes more patience.

Are paid Claude SEO skill packs worth it?

Sometimes, for niche platforms like Webflow. In general the free GitHub suites are larger, inspectable and more actively updated.

Is it safe to install skills from GitHub?

Treat them like any code. Read the SKILL.md and scripts, pin a version, and give connected servers the narrowest permissions possible.

Sources and methodology

Every tool in this guide was checked on 1 November 2026 against its primary source:

GitHub repositories (README, license, commands, install steps) for claude-seo, claude-blog, marketingskills, claude-skills, seomachine, seo-geo-claude-skills, awesome-claude-skills, mcp-gsc, DataForSEO MCP and Anthropic’s knowledge-work-plugins
Star counts and last push dates from the GitHub REST API on that date
Ahrefs MCP documentation at docs.ahrefs.com
Seller listings on Contra and Gumroad for paid packs
Community discussion on Reddit and skill directories (claudemarketplaces.com, claudskills.com, skills.sh) for common questions and objections

Star counts change daily, and repos rename commands between versions. If a command here does not work, check the repo’s README first.

Your next step

Pick the row in the stack table that matches your role. Install one suite, connect Search Console, and run a single audit or content brief on a page you know well. You will see quickly whether the output is specific to your site or generic. That one test tells you more than any list, including this one.

Claude Code Skills for SEO: The Complete 2026 Guide (Free and Paid)