The SEO Skill Benchmark
Which free SEO skill for Claude Code actually gives correct advice? seo-skill-bench runs every skill for real — headless, three times, one identical prompt — against a fixture site with 10 planted defects and 5 hallucination traps, then scores what each skill actually did against a machine-readable answer key. No vibes, no star counts: every transcript, score, and adjudication is public.
Watch: how the benchmark works and the full leaderboard, in two minutes.
| # | Skill | Composite | Detection | Trap avoidance | Judgment | Execution | Footprint | Resident | Median time | Cost/run | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | SEOAgent ↗ | 84.9 | 81% | 100% | 9.0/10 | 50% | light | 257t | 13 min | $4.46 | |
| 2 | claude-seo-skills (lhitches) ↗★ 1 | 76.0 | 52% | 100% | 8.0/10 | 100% | light | 131t | 7 min | $2.20 | |
| 3 | Agentic SEO Skill ↗★ 713 | 75.4 | 71% | 73% | 9.0/10 | 100% | light | 106t | 10 min | $2.81 | |
| 4 | Vanilla baseline (no skill) | 73.0 | 71% | 100% | 6.0/10 | 100% | — | — | 7 min | $2.31 | |
| 5 | claude-seo-skill (mangollc) ↗★ 36 | 70.4 | 67% | 73% | 7.0/10 | 100% | heavy | 1,315t | 9 min | $3.08 | |
| 6 | claude-seo ↗★ 10,215 | 67.7 | 38% | 100% | 6.0/10 | 100% | heavy | 2,687t | 7 min | $1.70 | |
| 7 | claude-seo-skills (lionkiii) ↗★ 18 | 66.6 | 52% | 73% | 6.0/10 | 100% | heavy | 4,848t | 12 min | $2.96 | |
| 8 | Marketing Skills (Corey Haines) ↗★ 36,047 | 66.6 | 52% | 73% | 7.0/10 | 100% | heavy | 8,770t | 8 min | $1.87 | |
| 9 | SEO/GEO Claude Skills ↗★ 20 | 66.0 | 57% | 100% | 6.0/10 | 100% | light | 20t | 9 min | $2.16 | |
| 10 | Distribb Skill ↗★ 87 | 65.9 | 43% | 100% | 7.0/10 | 100% | medium | 347t | 8 min | $2.33 |
Medians of 3 runs per skill on a pinned model (claude-sonnet-5), fixture pivot-saas. Composite = 40% defect detection + 25% trap avoidance + 25% blind judgment + 10% execution, pre-registered in RUBRIC.md before any runs. Observed fleet-to-fleet variance is roughly ±8 points — gaps inside that band are statistical ties. Footprint / Resident is context cost, deliberately outside the composite: an installed skill's description loads into the system prompt of every message, whether or not it fires — heavy resident footprints dilute the triggering of everything else installed (bands: light ≤280 tokens, medium ≤1,000, heavy >1,000; methodology in the benchmark README). Stars are GitHub stars at test time. SEOAgent maintains this benchmark and enters its own skill; every trap hit is hand-adjudicated with receipts in ADJUDICATIONS.md.
Looking for a Claude SEO checker? Read the board this way
A Claude SEO checker is a skill you install into Claude Code, Cursor, or Codex that audits your site from inside the repository that builds it. It crawls the live pages, reads the source that renders them, and reports what is actually broken — a canonical pointing at another origin, a sitemap that omits every published post, a blog index that only exists after JavaScript runs. It costs nothing beyond the model you already pay for, and because it can open the files it is complaining about, the output is a diff you review rather than a PDF you re-type.
That also makes checkers hard to compare. Every one of them will hand you a confident list of findings; the list is worthless if the findings are invented. So read the table above in this order:
- Trap avoidance first. Five things on the fixture are already correct. A checker that recommends “fixing” them will send you to edit working code on your own site. Anything below 100% here is a checker that hallucinates work.
- Detection second. The share of ten real, weighted defects the checker found. This is the number most people assume a checker is being scored on, and it is only useful once trap avoidance is clean.
- Judgment third. Blind-scored answers to open strategy questions — the part that separates a linter from an adviser.
- Footprint last, but do not skip it. An installed skill's description sits in the system prompt of every message you send, whether or not it fires. A heavy checker quietly degrades every other skill you have installed.
One thing no checker on this board does is track rankings. A checker inspects your pages and tells you what to fix; position and click data comes from Google Search Console, which some of these tools connect to separately. If you want the fuller argument for auditing from inside the repo, the write-up is in which free Claude SEO skill is best, and SEO tools for Claude Code covers the wider category.
How the benchmark works
Why you can trust a benchmark we maintain
You shouldn't trust it — you should check it. The rubric was pre-registered before any runs. Scoring is mechanical, against a manifest, with the scorer's self-tests in the repo. Every fleet is committed with full receipts: transcripts, scores, invalid runs, and hand-adjudications of every trap hit. That history includes the fleet where our own skill finished dead last — 9th of 9, below running Claude with no skill at all — and two scoring corrections that lowered our numbers, applied because they were correct. The benchmark caught a data-leak-class bug, a crawler auditing the wrong website, and a session-wedging hook in our product; fixing what it caught is why the skill now leads. Read the full story.
Reproduce it, or beat it
Everything is MIT-licensed: fixtures, manifests, harness, rubric, transcripts, adjudications. One command re-runs any entrant:
git clone https://github.com/aleclindz/seo-skill-bench.git && cd seo-skill-bench node harness/run.mjs --skill <id> --fixture pivot-saas --runs 3
Maintain a skill? Open an issue or PR adding it to skills.json — the next fleet picks it up and publishes your scores, good or bad. The top of this board is one good release away from changing hands. That pressure is the point. Before submitting, run your SKILL.md through the free Skill Grader — it scores the same context-footprint axes this board publishes (resident tokens, description honesty, body size) in seconds.
Show your score
Every entrant gets a live badge that re-renders from the current leaderboard — drop it in your README: 
[](https://seoagent.com/seo-skill-benchmark)
Want the skill at the top of the board?
SEOAgent is free for Claude Code, Cursor, and Codex — origin-verified crawls, evidence-cited findings, mechanical hallucination guards, and the field's only real migration planner.
FAQ
What is the best SEO skill for Claude Code?
On the current seo-skill-bench leaderboard, SEOAgent scores highest on the composite (defect detection, hallucination-trap avoidance, blind-judged strategy, execution). The benchmark is open source and every run is reproducible, so the answer can change with any release — check the live leaderboard.
What is a Claude SEO checker?
A Claude SEO checker is a skill you install into Claude Code (or Cursor or Codex) that audits a site from inside the repo it is built from: it crawls the live pages, reads the source that renders them, and reports what is wrong — missing canonicals, a sitemap that omits published posts, a client-rendered index, thin or contradictory metadata. Unlike a hosted SEO checker it needs no separate subscription and no crawl credits, because it runs on the model you already pay for and edits the files it is complaining about.
What is the best Claude SEO checker?
The leaderboard on this page is the answer, measured rather than asserted. Every free SEO skill for Claude Code we could find runs headlessly against the same fixture site with a known answer key, alongside a control row — Claude with no skill installed — that shows what the checker actually adds. Read the Detection column for how much of a real site's breakage a checker finds, and Trap avoidance for how often it invents a problem that is not there. A checker that scores well on detection but poorly on traps will hand you confident, wrong work.
Is a Claude SEO checker the same as a rank tracker?
No, and the benchmark does not score ranking data. A checker inspects your pages and tells you what to fix; a rank tracker records where you already place for a query. The skills on this board are checkers — they read your site and your repo. Position and click data comes from Google Search Console, which some of them can connect to separately.
How does seo-skill-bench score a skill?
Each skill runs headlessly, three times, with one identical prompt against a fixture site containing 10 planted defects and 5 traps (things that are already correct — recommending a "fix" is an objectively scored hallucination). Composite = 40% defect detection + 25% trap avoidance + 25% blind judgment + 10% execution, pre-registered before any runs. The published score is the median of the runs.
Can I trust a benchmark maintained by one of the entrants?
Check the receipts instead of trusting anyone: every transcript, score, invalid run, and adjudication is committed to the public repo — including the fleet where SEOAgent finished dead last, and scoring corrections that cut against it. The rubric is pre-registered, scoring is mechanical against a manifest, and one command re-runs any entrant.
How do I submit my SEO skill to the benchmark?
Open a pull request or issue on the seo-skill-bench GitHub repo adding your skill to skills.json (a GitHub repo with a SKILL.md, or an npm package). The next fleet run picks it up, and your scores — good or bad — are published with full receipts.