Claude Code SEO Skills
Claude Code SEO skills turn scattered checks into a repeatable run: parse a Search Console export, surface pages sitting just outside the top, verify meta tags and structured data, and build a brief for a specific query. The same checklist reaches every page instead of only the one somebody remembered. Technical SEO fits skills well precisely because it is made of checks with unambiguous answers: is there exactly one canonical on the page, does hreflang point both ways, is the title under sixty characters, does the article carry schema markup. People skip those out of inattention. A skill does not. A second benefit is that thresholds stop being a matter of taste: the skill holds them as numbers, so nobody has to argue about whether a description is long enough. It also keeps the run auditable: the same checklist produces the same report, so two people looking at one page reach the same conclusion.
The second area is data work. A thousand-row Search Console export cannot be read by eye — what matters in it only appears after grouping by position and by section. The skill records how to do that grouping and which thresholds are worth acting on.
The collection below covers audits, search data analysis and query-driven content prep.
Skills in this collection
83 skills in this collection; the 24 highest-rated are listed below.
- SEO-статья и кластеризация ключевых слов — SEO-контент и статьи is a Claude Code skill that builds a keyword-cluster table around a topic and writes a full SEO-optimised long-form article, delivering the output as an Excel workbook and a Markdown document. Reach for it when you need a new page to rank across a set of related queries, when an existing article gets little traffic and you can't pinpoint why, or when you want to cover a topic thoroughly without letting pages cannibalise each other's keywords. The skill starts by confirming target language and region, page type (blog post / product page / landing page), reader stage (awareness / comparison / purchase) and desired word count, then uses web search to collect related searches, People Also Ask questions and competitor titles, grouping them into 3–6 clusters by search intent. The Excel workbook has two sheets: keyword clusters with priority ratings, and a Competitor SERP sheet with tactics for outranking each result. The Markdown article includes three title candidates (up to 60 characters each), a meta description, H2 headings derived from cluster sub-questions, an FAQ section, internal-link suggestions and a cited-sources list. If a page on the topic already ranks, the skill recommends extending it rather than creating a new one to avoid cannibalization.
- Бриф для SEO-статьи по анализу SERP — Content Brief is a Claude Code skill that builds a detailed, SERP-driven content brief for a target keyword using live search results, competitor analysis, and the site's own data. Reach for it when you need to know exactly what an article must cover to rank for a specific query, when an editor needs a structured outline before handing work to a writer, or when you're deciding whether to create a new page or expand an existing one. The skill runs through the free Ryze MCP connector (`https://connector.get-ryze.ai/mcp`), added to Claude Code via `claude mcp add ryze --transport http https://connector.get-ryze.ai/mcp`, and supported in Claude.ai, Desktop, and Cursor as a custom connector. The workflow has five steps: pull the live Google top 10 via DataForSEO (or web search as fallback), skim the structure and angle of the top 3 pages, generate secondary keywords and question variants with volumes via Google Ads API, query Google Search Console for impressions the site already earns on the topic, then assemble the brief — target keyword plus secondaries with volumes, search intent and required format, proposed title and meta under character limits, H2/H3 outline with per-section notes, PAA questions, table stakes versus differentiating angle, 3–5 internal link sources with anchors, and a target word count matched to the ranking median.
- Анализ разрыва с конкурентом по SEO-запросам — Competitor Gap is a Claude Code skill that identifies keywords where a competitor ranks in the top 20 while your site receives zero impressions, then prioritizes them by search volume and realistic chance of winning. Reach for it when you want to know why a rival outranks you, need a ready list of topics for new content, or want to find existing pages worth improving to close the gap. The skill runs through the free Ryze MCP connector (`https://connector.get-ryze.ai/mcp`), added with `claude mcp add ryze --transport http https://connector.get-ryze.ai/mcp`; competitor rankings come from DataForSEO Labs (or Ahrefs/Semrush if connected in the workspace), while your real data comes from Google Search Console over the last 90 days. Results are split into three buckets: Gap (competitor top 20, you have zero impressions — create new content), Behind (both ranking but competitor is 5+ spots ahead — improve your page), and Ahead (your existing moat). Competitor metrics are estimates; yours are real — the skill labels each explicitly. Branded competitor terms are excluded by default but can be surfaced separately as "vs / alternative" page candidates. This skill compares an external competitor against your site, not a technical SEO audit of your own domain.
- Анализ трафика из AI-поисковиков через GA4 — AI Visibility is a Claude Code skill that measures real traffic from AI search engines — ChatGPT, Perplexity, Claude, Gemini, Copilot — using actual Google Analytics 4 referral data, and identifies which pages those engines actively cite. Reach for it when you need to know whether ChatGPT is actually sending visitors to your site, when a client asks for an AEO or GEO audit, when you want to find out why competitors appear in AI answers and you don't, or when you need to prove AI-channel ROI with conversion data. The skill connects through the free Ryze MCP connector (one command: claude mcp add ryze --transport http https://connector.get-ryze.ai/mcp) and calls GA4 tools to pull sessions by AI source and period, daily trend, cited landing pages with conversions, and referral-level detail. It then cross-references Google Search Console to surface pages with strong organic queries but zero AI referrals — the restructuring shortlist. Output includes an AI traffic verdict, engine and landing-page tables, a citation recipe derived from your top-cited pages, and a concrete top-5 action list. Unlike tools that sample prompts and estimate visibility, this skill reads the clicks that actually happened. Requires GA4 to be set up; a sessionSource-filter fallback is available if the dedicated AI-traffic tools aren't in the workspace.
- Отслеживание позиций в поиске через Google Search Console — Rank Tracking is a Claude Code skill that compares keyword positions between any two time windows using real Google Search Console data — no paid rank-tracker subscription needed. Reach for it when you want to know what happened to your rankings after a Google algorithm update, when you need to see which pages gained or lost ground over the past month, or when you want hard numbers on whether your SEO work is paying off. The skill requires the free Ryze MCP connector (`https://connector.get-ryze.ai/mcp`), added to Claude Code with a single command. Under the hood it pulls two periods of Search Console data (default: last 28 days vs. the previous 28), joins query- and page-level rows, and classifies results into Winners, Losers, New, and Lost — sorted by clicks impact rather than raw position delta, so high-value queries surface first. For top losers, it optionally checks the live SERP via DataForSEO if that tool is available in the workspace. Output leads with a plain-language verdict, followed by four tables. This is retrospective change analysis, not real-time daily monitoring. GSC data lags roughly two days, so today is always excluded. Works for any site connected to Google Search Console; not applicable without it or without access to the Ryze connector.
- Анализ ссылочного профиля и link gap — Backlink Profile Analysis is a Claude Code skill that audits the inbound link profile of any domain — referring domains, authority distribution, anchor breakdown, new and lost links — and, when a competitor is named, computes the link gap and produces an outreach-ready list of domains linking to them but not to you. Reach for it when you want to understand who links to your site and whether the profile looks healthy; when you need a side-by-side comparison with a competitor to spot untapped link sources; or when an SEO audit calls for flagging spammy anchor patterns, link velocity spikes, or low-authority domain dominance. Built on the free Ryze MCP connector (`https://connector.get-ryze.ai/mcp`), added to Claude Code with a single command. Primary data source is DataForSEO Backlinks (summary, referring_domains, anchors); fallback providers are Ahrefs or SEMrush; if none are connected, the skill stops and names what's missing — it never fabricates link counts. An optional cross-check compares linking domains against referral sessions in Google Analytics (own site only). The skill covers inbound external links, not internal link structure. All third-party figures are labeled as index estimates. Works for any publicly indexed domain; requires at least one data provider connected.
- SEO-аудит сайта через Search Console и GA4 — SEO-аудит сайта is a Claude Code skill that runs a full SEO audit of a website using its own live data from Google Search Console and GA4 via the Ryze MCP connector — no third-party estimates, only real numbers. Reach for it when organic traffic has dropped and you can't tell why, when you want to find pages one push away from page one, when you need to diagnose indexation issues, or when the site ranks but nobody clicks. The skill connects through the free Ryze MCP (`https://connector.get-ryze.ai/mcp`), added with a single command in Claude Code and compatible with claude.ai, Desktop, and Cursor. It compares the last 90 days against the prior 90 across clicks, impressions, CTR, and positions; flags decaying pages (clicks down >30%), CTR anomalies (position ≤5, CTR <2%), striking-distance queries (positions 8–20 with high impressions), and content cannibalization. GA4 data is cross-checked to surface pages that hold rankings but lose revenue, and AI referral traffic is reported separately. Output is a verdict paragraph, then structured tables for Indexation, Losers, Quick Wins, and Cannibalization, closing with a prioritized five-item action list. This skill diagnoses and prioritizes — it does not crawl site structure or audit page speed (separate tools cover those).
- SEO-исследование ключевых слов и контент-стратегия — Keyword Research is a Claude Code skill that runs a full keyword research cycle — from a seed topic to a prioritized content plan with real search volumes, CPC data, and topic clusters. Reach for it when you need to figure out what to rank for, when a flat keyword list isn't enough and you want intent-grouped clusters with actionable page priorities, or when you're planning Google Ads campaigns and need traffic-value signals alongside organic data. The workflow has four steps: seed expansion via Google Ads Keyword Planner (up to ~200 ideas), historical metrics pull (monthly volume, competition, CPC, 12-month trend), optional DataForSEO enrichment for the top ~50 keywords (search intent, keyword difficulty, SERP features — silently skipped if unavailable), and a Google Search Console check to flag queries where the site already holds positions. Output is per-cluster tables (target keyword, supporting keywords, volume, CPC, intent, current position or "new page") plus a ranked attack plan weighted by volume and winnability. Requires the free Ryze MCP connector; DataForSEO integration is optional. The skill handles research and planning, not on-page optimization or technical audits. CPC data doubles as a negative-keyword source for Google Ads. Works exclusively with Google data sources — not suitable if Bing or third-party volume providers are required.
- SEO-автоматизация: ключевые слова и анализ конкурентов — seo-ops is a Claude Code skill that automates SEO operations: keyword research, competitor content gap analysis, Google Search Console auditing, and trend detection. It includes four scripts — content_attack_brief.py for the full keyword intelligence pipeline including BOFU keywords and decaying pages, gsc_client.py as a GSC API client, gsc_auth.py for one-time OAuth setup, and trend_scout.py for multi-source trend monitoring across Google Trends, HN, Reddit, and X. Every keyword is scored by Impact × Confidence (up to 100), keeping focus on the highest-value targets first. The skill suits SEO specialists and content marketers who need to surface striking-distance keywords, spot pages with declining traffic, or build a prioritized content brief from scratch.
- GEO-аудит сайта для AI-поисковиков — geo is a Claude Code skill that performs full GEO-first audits to optimize websites for AI-powered search engines — ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews — while preserving traditional SEO foundations. It orchestrates up to 5 parallel subagents covering citability scoring, AI crawler access and llms.txt generation, brand mention scanning across Reddit, YouTube, Wikipedia, and LinkedIn, Schema markup validation and generation, technical SEO, and E-E-A-T content quality assessment across 14 specialized components. A composite GEO Score (0–100) is calculated from six weighted categories, and the `/geo report-pdf` command produces a client-ready PDF with charts. Designed for agencies, SEO professionals, and SaaS, e-commerce, or local businesses that need to capture traffic from the growing wave of AI-referred sessions.
- Анализ упоминаний бренда для AI-поиска — geo-brand-mentions is a Claude Code skill that scans brand mentions across key platforms and computes a Brand Authority Score (0–100) to measure a brand's visibility in AI search engines and large language models. An Ahrefs study published in December 2025, covering 75,000 brands, found that unlinked brand mentions correlate roughly 3× more strongly with AI citations than traditional backlinks, and the platform where a mention appears determines its weight significantly. The skill evaluates presence on YouTube (correlation ~0.737, the strongest signal), Reddit, Wikipedia, Quora, LinkedIn, and other platforms that AI models rely on for entity recognition and citation decisions. Each platform receives a scored breakdown with explicit criteria and actionable recommendations for improving standing. It is designed for marketers and SEO professionals pursuing GEO strategies who need to understand why ChatGPT, Perplexity, or Google AI Overviews fail to mention their brand.
- SEO-аудит: краул и скорость сайта в Excel — pull-crawl-and-site-speed is a Claude Code skill that merges a domain's most recent nightly crawl and nightly site speed audit from Tapa's monitoring pipelines into a single styled Excel workbook containing a Summary tab, Crawl Pages, Speed Pages, and 30-day trend charts for both datasets. It operates via the Tether MCP with a single synchronous `tapa_nr_run` call using `report: "both"` — no polling or job IDs required. Claude renders KPI cards and charts directly in chat, then provides a download link for the workbook and a deep link into the Tapa Crawler Dashboard with the domain preselected. The skill closes with 4–8 cross-referenced AI insights connecting crawl health and speed data — heavy pages versus slow templates, health trends versus performance trends — citing real numbers and URLs from the returned data. Built for SEO teams and technical auditors who need the full nightly picture of a site's health in a single request.
- Семантическое сопоставление URL при миграции сайта — tapa-redirect-mapper is a Claude Code skill that builds a site-migration redirect map by matching every old-site URL to its top-3 best new-site equivalents, ranked by content similarity across page titles, meta descriptions, and H1s using embedding-based scoring. It requires two Screaming Frog crawl exports — one for the old site, one for the new — in .csv or .xlsx format with columns Address, Title 1, Meta Description 1, H1-1, and Indexability. Files are staged via tapa_rm_upload, the job submitted through tapa_rm_run, and results polled with tapa_rm_result until complete. Delivery includes a downloadable workbook (Summary, Redirect Map, and Best Match Only sheets) plus an in-chat visualisation of KPI cards and charts rendered as a self-contained HTML artifact with inline SVG. The tapa_rm_* tools run over the Tether MCP; access is limited to a test allow-list while Tapa Skills are in early testing. Ideal for SEO teams and developers who need to map hundreds of URLs programmatically before a site launch.
- Аудит скорости сайта из Tapa в Excel-отчёт — pull-site-speed is a Claude Code skill that retrieves the most recent nightly site speed audit for a domain from Tapa's monitoring pipeline via the Tether MCP connector. It returns a styled Excel workbook with three tabs — Summary (run timestamps, Lighthouse lab scores, CrUX field data, carbon per view), Pages (every measured URL, mobile and desktop), and Trends (30-day line charts of homepage performance) — along with a deep link to the Tapa Crawler Dashboard's Site speed view. The skill also renders an in-chat visual of all KPI cards and charts, and delivers 3–6 AI insights citing real numbers: slow page types, lab-vs-field discrepancies judged against Core Web Vitals thresholds, trend direction, and the single most valuable fix. Built for SEO and performance teams who need stored nightly data; for a fresh live measurement use tapa-full-site-speed-audit or tapa-page-speed-audit instead.
- Поиск каннибализации страниц по данным Screaming Frog — tapa-cannibalisation-detector is a Claude Code skill that scans an entire site for keyword and content cannibalisation using Tapa's standalone Cannibalisation Detector via the Tether MCP. It accepts a whole-site Screaming Frog crawl export (.csv or .xlsx) containing Address, Title 1, Meta Description 1, H1-1, and Indexability columns, then compares every pair of indexable pages and flags those with ≥70% similarity across titles, meta descriptions, and H1s. Findings are visualised directly in chat as KPI cards and charts rendered from real result data, alongside a downloadable workbook with Summary and Cannibalisation Risks sheets. This is the whole-crawl detector — distinct from Content EEATer's Search Console cannibalisation sub-check — making it the right choice for SEO auditors who need to surface competing and overlapping pages across a full site crawl.
- Аудит внутренней перелинковки для новых страниц — tapa-internal-link-audit is a Claude Code skill that finds the best internal-link sources for many new pages at once, ranking a site's existing pages by content similarity to each new page. It requires two Screaming Frog exports — a full-site crawl and a list of new or planned pages — in .csv or .xlsx format, with columns Address, Title 1, Meta Description 1, H1-1, and Indexability. Using Tether MCP tools (`tapa_ila_upload`, `tapa_ila_run`, `tapa_ila_result`), the skill processes both files and produces a workbook with a Summary sheet plus one ranked source list per new page. Top opportunities are always visualised in chat as KPI cards and charts; a download link to the expiring workbook is included separately. Designed for content plans, site migrations, and whole-category rollouts — for a single page, `/tapa-content-eeater` covers link-source ranking as part of its own audit.
- Tapa — SEO-отчёт по краулу сайта — pull-crawl is a Claude Code skill that retrieves the most recent nightly crawl of a domain from Tapa's monitoring pipeline via the Tether MCP and delivers a structured SEO audit report. The output includes a styled Excel workbook with three sheets — Summary (run timestamps and health KPIs), Pages (every crawled page with its issues), and Trends (30-day crawl-health line charts backed by a hidden data sheet) — along with a download link and a deep link into the Tapa Crawler Dashboard. In chat, Claude renders all KPI cards and charts as an HTML artifact or interactive widget, then adds 3–6 data-driven insights covering broken pages, indexability shifts, issue hot-spots, and the single highest-value fix. This skill is the right choice when the user wants last night's or the latest stored crawl data; for an on-demand fresh crawl, use tapa-site-crawler instead.
- Краулинг сайта и SEO-аудит страниц — tapa-site-crawler is a Claude Code skill that crawls a website via the Tether MCP and delivers a downloadable CSV containing URLs, status codes, titles, on-page fields, and a `page_category` label for every page — Homepage, PDP, PLP, Article, Service, About, Contact, or Other. A single crawl covers up to 10,000 pages; specifying the site's CMS (Shopify, WordPress, Magento, Squarespace, and others) sharpens the page-type classification using platform-specific URL patterns. The resulting file is the natural input for `/tapa-cannibalisation-detector`, `/tapa-redirect-mapper`, and `/tapa-internal-link-audit`. Designed for SEO professionals and digital teams who need a fresh crawl file for technical audits, redirect mapping, or internal link analysis.
- Анализ угасания трафика по запросам Search Console — tapa-content-decay-detector is a Claude Code skill that compares two Google Search Console Queries exports and classifies every search query as Declining, Stable, or Growing with click, impression, CTR, and position deltas. It operates through the Tether MCP using four dedicated tools — tapa_cdd_options, tapa_cdd_upload, tapa_cdd_run, and tapa_cdd_result — and accepts .csv or .xlsx files, with Period 1 as the older window and Period 2 as the newer one. Once the analysis completes, it delivers a downloadable workbook (Summary, Full List, and per-status sheets) alongside an in-chat visualisation of KPI cards and charts rendered as a self-contained HTML artifact. Built for SEO specialists and content teams tracking traffic decline and wanting to prioritise pages that need refresh or recovery.
- Массовый перевод метаданных сайта для международного SEO — tapa-auto-translate is a Claude Code skill that translates a site's page titles, H1s, and meta descriptions into other languages using Google Cloud Translation applied to a Screaming Frog crawl file via Tether MCP tools. It accepts a CSV or XLSX export with Address, Title 1, Meta Description 1, and H1-1 columns, then calls `tapa_at_upload`, `tapa_at_run`, and `tapa_at_result` in sequence, producing an Excel workbook with one tab per target language chosen from a catalogue of 133 options. Results appear in chat as KPI stat cards showing pages translated and languages covered, alongside a clickable download link to the workbook. It's built for SEO teams who need to localise site metadata at scale for international search without manually translating each page.
- Google Search Console — SEO-аналитика и управление сайтами — gog-searchconsole is a Claude Code skill that performs Google Search Console operations through the gog command-line tool. It exposes four commands: `query` and `searchanalytics` for running search analytics queries, `sitemaps` for listing, retrieving, submitting, and deleting sitemaps, and `sites` for listing and inspecting verified Search Console properties. Authentication is verified upfront with `gog auth list`, while `--readonly` and `--dry-run` flags prevent unintended mutations. The skill is aimed at SEO specialists and developers who need to automate search visibility monitoring, indexation checks, and sitemap management without manually navigating the Google Search Console interface.
- Комплексный аудит скорости сайта с Lighthouse и Core Web Vitals — tapa-full-site-speed-audit is a Claude Code skill that runs a comprehensive speed audit across an entire site via the Tether MCP using Tapa's toolset. It crawls the domain, benchmarks Lighthouse scores by page type (Homepage, PDP, PLP, Article, Service, and more) on mobile and desktop, pulls real-user Core Web Vitals for every country available in the Chrome UX Report, estimates CO₂e per page view, and generates CMS-specific recommendations for platforms like Shopify, WordPress, and Magento. Up to three competitor domains can be benchmarked side by side. Results are visualised in chat with KPI cards and data charts, and a downloadable workbook is provided with per-page results and field data by country. A typical run takes 10–30 minutes, making this the right choice for SEO professionals and developers who need a full-site performance picture rather than spot-checking individual URLs.
- GEO+SEO аудит видимости сайта в ИИ-поиске — geo-audit is a Claude Code skill that runs a full GEO+SEO audit of any website by orchestrating five specialized subagents in parallel. The subagents cover AI citability scoring, platform-specific readiness for ChatGPT, Perplexity, Google AI Overviews, Gemini, and Bing Copilot, technical infrastructure including robots.txt and llms.txt access for AI crawlers, content E-E-A-T quality, and Schema.org structured data validation. Results are aggregated into a weighted composite GEO Score (0–100) across six categories — AI citability at 25%, brand authority at 20%, E-E-A-T at 20%, technical GEO at 15%, schema at 10%, and platform optimization — paired with a prioritized action plan. Built for teams who need to measure and improve their site's visibility in AI-generated responses and generative search engines.
- SEO-анализ данных краулинга с рекомендациями — data-analyst is a Claude Code skill that processes SEO crawl data, detects patterns, and generates actionable insights for the SEOcrawler V2 project. It queries Supabase to aggregate metrics across crawls and performs statistical analysis — mean, median, standard deviation, correlations, and time series — then produces visualizations with matplotlib/seaborn exported as PNG or SVG. Coverage spans crawler performance metrics, meta tag completeness, Core Web Vitals distributions, mobile responsiveness scores, and Dutch market compliance checks such as KvK/BTW presence. Competitive analysis includes SERP position correlations and market segment benchmarks. Deliverables follow a structured report format with an executive summary and recommendations prioritized into immediate, 30-day, and quarterly actions, all backed by reproducible analysis code meeting statistical significance thresholds (p < 0.05, n > 30).
FAQ
Can Claude Code replace an SEO specialist?
No. Skills cover the reproducible half of the work: checks, export analysis, briefs from a template. Deciding which queries are worth pursuing, what content to write and how to differ from competitors stays with a person — that takes market knowledge, not checklist execution.
Do I need Search Console access for these skills to be useful?
Not necessarily. Some skills work from a downloaded report file, others work against the live pages, checking markup and meta tags. Direct API access is only required by skills that pull fresh data themselves.
All Claude Code skills