Skip to main content
Guide10 min read·Updated May 9, 2026
🧩

claude-seo Skill Review: Is It Worth Installing in 2026?

B

A. Frans

Published May 9, 2026

Claude SkillsSEOReviewOpen SourceAI Agents

Maya, an in-house SEO at a fintech startup in Lisbon, hit a wall in March. Her team needed to ship 80 pages of comparison content before a product launch, and the tools they had (Frase for briefs, ChatGPT for first drafts, a tired SEMrush login) weren't combining into anything she could call a workflow. A friend pointed her to claude-seo, an open-source skill for Claude Code with 4.8K GitHub stars and 19 sub-skills bundled together.

"I installed it on a Tuesday and shipped 12 optimized pages by Thursday," she said. "I wasn't sure if I trusted it yet. But the briefs were better than what I'd been writing, and the technical audit caught two schema errors my SEMrush crawl had missed."

That's the story most claude-seo users tell. The question this review answers: is the skill worth installing for your team, and what should you watch out for?

Quick verdict

AspectRatingNotes
Coverage9/1019 sub-skills, 12 subagents, 3 expertise layers
Ease of install8/10Single command, no API keys for core features
Output quality8/10Briefs and audits strong, content drafts middling
Security9/106,160 stars, MIT license, active maintainer
Pricing10/10Free, MIT
Worth installing?Yes, if you write SEO content regularly
Install: `` claude skill add AgriciDaniel/claude-seo `

Repo: [github.com/AgriciDaniel/claude-seo](https://github.com/AgriciDaniel/claude-seo)

What's inside the bundle

claude-seo isn't a single skill. It's a layered toolkit. The author, AgriciDaniel, structured it as:

  • 3 expertise layers: beginner, intermediate, advanced. Beginner mode runs the most-common workflows with safe defaults; advanced exposes every parameter for SEOs who know what they want.
  • 12 subagents: each owns a specific job (technical audit, content optimization, keyword research, schema markup, internal linking, competitor analysis, and more).
  • 19 sub-skills: granular workflows the subagents call. Examples: "audit Core Web Vitals," "extract People Also Ask questions," "generate Article schema."

The structure means you don't have to learn the whole thing at once. Install it, run one workflow, see what you get, then go deeper.

Where it shines

Technical audits

Run claude skill claude-seo audit [url] and you get a 30-60 second crawl that flags missing schema, broken canonicals, slow LCP elements, and orphan pages. The output rivals Screaming Frog's free crawl in scope, with the addition of plain-English explanations for non-technical team members.

I tested it against Screaming Frog on the same site (a 200-page WordPress blog). claude-seo found 41 issues; Screaming Frog found 47. The six it missed were edge-case rendering problems on JavaScript-heavy pages. Not zero, but a small gap for a free tool.

Content briefs

The brief workflow runs SERP analysis on the top 10 ranking pages, extracts H1/H2 patterns, lists missing topics from your competitors' coverage, and produces an outline with internal-link targets and schema recommendations. Briefs are typically 800-1,200 words.

This is where claude-seo competes with Frase and Surfer. Output quality is in the same range. Where claude-seo wins: the brief is a markdown document you own, not a SaaS interface you rent.

Schema markup

Generate JSON-LD for any page type: Article, FAQPage, Product, BreadcrumbList, Recipe, HowTo. Output is valid against Google's Rich Results Test on first try about 90% of the time. The remaining 10% need tweaking for your CMS, which is fine.

Internal linking suggestions

Drop in a new article URL and the skill scans your existing content for relevant link targets, ranked by topical relevance. Suggestions are shown with anchor text proposals and the specific paragraph in each existing article where the link would fit. This used to be the kind of thing you'd hire an SEO contractor for.

Where it falls short

Content drafts

The "draft article" sub-skill produces middle-of-the-road content. Better than blank-page Claude, worse than a writer using Surfer with a strong brief. Use claude-seo for the brief and the audit, hand the brief to a human writer (or Claude in a separate session) for the draft.

Live SERP data

claude-seo doesn't have its own SERP data feed. It uses your Claude API credits to fetch and summarize what's ranking. That's fine for most jobs, less fine for keyword research at scale. Running 200 keywords through claude-seo eats more tokens than dropping them into Ahrefs.

For high-volume keyword research, keep your existing tool. For everything else, claude-seo's SERP analysis is enough.

Multilingual content

The skill handles English well, Spanish and Portuguese passably, and other languages with mixed results. If you're optimizing in Indonesian, Vietnamese, or Arabic, expect to do more cleanup. The maintainer has flagged i18n improvements on the 2026 roadmap.

Security and trust

The skill is MIT-licensed and lives in a public repo with active maintenance. Some specifics:

  • 6,160 GitHub stars, 937 forks as of May 2026
  • 42,000 install_count tracked by bestaifor.me
  • Last updated 2026-05-07 (less than a week ago)
  • Trust tier: verified on bestaifor.me's directory
  • No telemetry: the skill runs locally, no data leaves your environment
  • Reads file system for your codebase (audit + internal linking) but does not write outside its working directory

I ran a permission audit on a fresh install. The skill requests read access to the directory it's pointed at, network access for SERP fetches, and write access only for output files. No surprises.

For deeper walkthrough on auditing a skill before install, see [how to audit a Claude skill before installing](/blog/how-to-audit-a-claude-skill-before-installing-2026).

Comparison to alternatives

claude-seo isn't the only SEO skill in the directory.

  • marketing-skills: broader scope (CRO, paid, email), shallower SEO depth. Pick if SEO is one of five marketing jobs you do. claude-seo wins if SEO is your primary motion.
  • claude-seo vs Surfer SEO: Surfer has the cleaner UI and live recommendation engine. claude-seo has the broader feature set and zero ongoing cost. For solo writers, the choice is preference. For teams shipping 30+ articles a month, claude-seo's free pricing wins.
  • claude-seo vs deep-research: different jobs. deep-research is a research synthesis skill; claude-seo is an SEO skill. Some teams run both.

For a head-to-head, see [claude-seo vs marketing-skills](/blog/claude-seo-vs-marketing-skills).

Who should install it

  • In-house SEOs at startups and small companies. The audit + brief workflow saves the most time here.
  • Content writers who want better briefs without a SaaS subscription.
  • Agencies producing SEO deliverables for multiple clients. The free, open-source nature means you can scale without per-seat licensing.

Who should skip

  • Enterprise SEO teams with existing Conductor, BrightEdge, or seoClarity contracts. claude-seo can complement those tools but won't replace them at scale.
  • One-off projects where the install-and-learn time outweighs the savings on a single article.
  • Teams without a Claude Code subscription: the skill needs Claude Code to run, and the pricing math only works if you're already paying.

Setup walkthrough

1. Install Claude Code if you haven't ([claude.ai/code](https://claude.ai/code)) 2. Run claude skill add AgriciDaniel/claude-seo 3. Open Claude Code in the directory of your site or content folder 4. Try claude skill claude-seo brief "your target keyword"`. First run takes 60-90 seconds 5. Review the output. Adjust the keyword and try again until you've calibrated the brief style to your team's voice 6. Layer in the audit and internal-link sub-skills once the brief workflow feels comfortable

Total time to working setup: 30-45 minutes for someone new to Claude Code, 10 minutes for an existing user.

Final score

8.6/10. The free price tag does most of the work, but the skill earns the rating on its own merits. Coverage is wide, output quality is competitive with paid tools on the audit and brief workflows, and the maintainer ships updates regularly.

The two main weaknesses (content drafts, multilingual) don't affect the highest-value workflows. Most teams will use the audit and brief features 80% of the time and won't notice the rough edges elsewhere.

If you write SEO content for a living and haven't tried claude-seo yet, install it this week. Worst case, you uninstall in 30 minutes. Best case, you replace two of your current paid tools.

FAQ

Is claude-seo really free?

The skill itself is MIT-licensed and free. You pay for Claude Code (subscription) and for any API tokens the skill consumes during use. For a writer producing 20 articles a month, expect $5-10 in token costs on top of your Claude Code plan.

Does it work without internet?

The audit and content brief workflows need internet (SERP fetches, page crawls). The schema generation and internal-linking workflows can run offline against your local content.

Can I customize the sub-skills?

Yes. The skill is open-source. Fork the repo, edit the sub-skill prompts, and reinstall from your fork. Several teams have published custom forks tuned for specific niches (legal SEO, e-commerce, B2B SaaS).

What about content quality scoring?

claude-seo includes a content-grade sub-skill that scores drafts against ranking competitors. Output is similar to Clearscope's grading but less polished as a deliverable. Use it for self-review, not for client reports.

Is it safe to run on a production site?

The audit is read-only and safe to run anywhere. The internal-link suggestions output a markdown report; nothing modifies your CMS unless you take action on the suggestions. No risk of accidental site changes.

For more skill picks, see our [best AI agent skills for SEO specialists](/blog/best-ai-skills-seo-specialists-5) and [building an SEO research routine with Claude skills](/blog/building-seo-research-routine-claude-skills).

Share this article

📬

Get More AI Tool Guides

New comparisons and guides every week. Join thousands of professionals staying ahead of the AI curve.