AI NewsOpen sourceAnnouncement
Eight open-source SEO skills connect keyword research to a site's real data
A new MIT-licensed set of eight SEO and GEO skills for Claude runs keyword research, rank tracking and AI-visibility checks against a site's own Search Console, GA4 and Google Ads data.
Image: GitHub
Why it mattersA team can stop paying $120 a month for rank estimates and instead read the real positions Google already records for them, from inside the coding agent.
Ryze AI has published open-seo-mcp-skills, an MIT-licensed set of eight SEO and generative-engine-optimisation skills for Claude that pull data from a site's own Google Search Console, GA4 and Google Ads accounts through an MCP connector. The repository went up on 29 August 2026 and has 255 stars a week later.
The pitch in the README is direct. Most "open-source SEO tools" are a user interface over the DataForSEO API, and the code being free does not make the data free: either the user brings a DataForSEO key and pays per request, or a hosted product adds a markup on every data call. Ryze puts the number at around 28 percent, according to the repository. These skills take a different approach by running against accounts the site owner already controls.
What the eight skills cover
The set includes seo-audit, which reads Search Console and GA4 for indexation state, click-through anomalies and decaying pages, keyword-research, which combines Google Ads keyword planner numbers with DataForSEO for ideas and clustering, rank-tracking, which computes position movers between any two periods from GSC alone with no third-party tracker, competitor-gap, backlink-check, content-brief, and seo-vs-ads, which compares Google Ads spend against organic ranking for the same query.
ai-visibility is the one that applies directly to the GEO problem. It reads GA4's AI-referral traffic to answer which engines are sending real users to the site and which pages they choose to cite. That is the measured version of the question every SEO team is guessing at right now.
What connects to what
The skills are provider-neutral Markdown that works with any MCP the host connects. The default path uses the Ryze MCP connector, which handles OAuth for Search Console, GA4 and Google Ads once and bundles DataForSEO through the workspace so no separate API key is needed. The README says Ahrefs and Semrush can also be connected if the site owner already pays for them.
The tools inside are namespaced by provider, for example google_search_console__runRawSearchAnalytics and google_ads__generateKeywordIdeas. A skill author who is unsure of a request shape can call native__get_provider_docs to fetch the official API docs on demand.
An in-house developer or a small marketing team pays somewhere between $120 and $500 a month for a rank tracker that guesses at data the site already owns, because the real data can only be reached through an OAuth login that the tool company cannot automate. The pattern here is the interesting part beyond this specific project: a Markdown skill plus an MCP connector to the account is enough to replace a category of hosted tools that exist because the OAuth was hard to package.
The two caveats to state plainly. First, the free skills route through a hosted MCP for the connector layer, so a team that wants no third party at all is still writing the OAuth code themselves. Second, DataForSEO is a paid API and its costs still show up on the site's own DataForSEO account when the skills request competitor keywords or SERPs. The project removes the markup, and the underlying data bill remains.
Source
This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.
Get AI News in your inbox
New developer tools, model and agent releases, and how teams are actually using them to release software. Short, and only when there is something worth reading.
