Skip to content
Get packs

Guide · Agent skill

How to use the Google SEO APIs agent skill (seo-google)

Updated · written with AI from the project's own documentation and checked against our catalogue data

The seo-google skill gives an AI agent direct access to Google's own SEO data: Chrome field data for Core Web Vitals, URL indexation status, Search Console search performance, and organic traffic from Google Analytics 4. This guide is for SEO specialists, site owners and agencies who may never have used an agent skill before. By the end you will know what the skill does, how to install it, how to set up the credentials it needs, and how to ask for results in plain language.

Google SEO APIsAgriciDaniel/claude-seo · MIT

Quick start

  1. 1Download and unzip. Download the seo-google ZIP from the directory page and unzip it.
  2. 2Place the folder. Move the seo-google folder into ~/.claude/skills/ for personal use or into .claude/skills/ in your project. In the Claude apps, upload it under Settings, then Capabilities, then Skills.
  3. 3Create Google credentials. In a Google Cloud project, create an API key and/or service account, and grant access to your Search Console property.
  4. 4Save the config file. Create ~/.config/claude-seo/google-api.json with your API key or service account path, default property and GA4 property ID if you use GA4.
  5. 5Run the credential check. Ask the agent to check credentials. It runs google_auth.py and reports your credential tier.
  6. 6Make a first request. Ask for a real data check, such as CrUX field data for a URL or Search Console clicks for a property.

What the seo-google skill makes the agent do

An agent skill is a folder of instructions and helper files that an AI agent loads when your request matches the skill's description. This one teaches the agent to query Google's SEO services instead of relying only on crawl-based guesses.

The skill covers Search Console (Search Analytics, URL Inspection and Sitemaps), PageSpeed Insights v5, CrUX field data with 25-week history, the Indexing API v3, and GA4 organic traffic. It runs helper scripts and uses API credentials you save locally.

  • Core Web Vitals: real Chrome user metrics (CrUX) alongside Lighthouse lab scores
  • Indexation: whether Google has indexed a URL, with verdict, coverage state and canonical
  • Search performance: clicks, impressions, CTR and position from Search Console
  • Organic traffic: GA4 reports for organic search and top landing pages

Install the seo-google skill with a ZIP file

Download the ZIP for the skill from the directory page, then unzip it. Put the resulting seo-google folder in one of two places: your personal skills folder at ~/.claude/skills/, or a project-specific .claude/skills/ folder inside the project you work in.

If you use the Claude apps instead of Claude Code, you can upload the skill under Settings, then Capabilities, then Skills.

  • Personal install: ~/.claude/skills/seo-google
  • Project install: .claude/skills/seo-google inside your project folder
  • Claude apps: Settings → Capabilities → Skills, then upload the skill

Set up Google credentials before running commands

Almost every command needs credentials. The skill checks them first with its google_auth.py helper. Credentials are stored in a local config file at ~/.config/claude-seo/google-api.json. The skill needs Python for its bundled scripts.

Setup needs a Google Cloud project with an API key and/or a service account, plus Search Console access to the property you want to check. GA4 commands also need a GA4 property ID. Keyword and volume commands need a Google Ads developer token and customer ID. The skill's setup command, /seo google setup, gives step-by-step instructions.

  • Config file: ~/.config/claude-seo/google-api.json
  • Fields include service_account_path, api_key, default_property, and ga4_property_id
  • Credential tiers: Tier 0 (API key) runs PageSpeed, CrUX, CrUX history, YouTube and NLP; Tier 1 adds Search Console (gsc, inspect, sitemaps) and indexing; Tier 2 adds GA4; Tier 3 adds Keyword Planner commands
  • The agent reports which tier it detected before running commands
"${CLAUDE_PLUGIN_ROOT}/scripts/claude-seo" run google_auth.py --check --json

Trigger the skill with natural requests

You do not need to memorise command names. The skill loads when your request includes phrases such as search console, GSC, PageSpeed, CrUX, field data, indexing API, GA4 organic, URL inspection or real CWV data.

You can also call commands directly with the /seo google prefix, using a URL or property as the argument.

  • Ask: Check the real field data for CWV on example.com pages
  • Ask: Show me 25-week Core Web Vitals trends for this URL
  • Ask: Is this URL indexed in Google? Run URL inspection on it
  • Ask: Pull my Search Console clicks and impressions for the last 28 days
/seo google crux-history <url>

Example commands and what they return

Commands take a URL or a Search Console property. The output is saved as data the agent can read and summarise, and the skill includes report templates for Core Web Vitals audits, Search Console performance and indexation status.

The CrUX history command shows whether metrics are improving, stable or degrading, with per-metric trend direction, percentage change and weekly p75 values. The Search Console command includes quick-win detection for queries ranking at positions 4 to 10 with high impressions.

  • /seo google pagespeed <url>: Lighthouse lab data plus CrUX field data, mobile and desktop
  • /seo google crux <url>: field data only, faster, no Lighthouse run
  • /seo google gsc <property>: clicks, impressions, CTR and position for the last 28 days
  • /seo google inspect <url>: URL Inspection verdict (PASS or FAIL), coverage state and robots.txt status
  • /seo google ga4 [property-id]: GA4 organic traffic report
  • /seo google quotas: rate limits for all APIs
/seo google gsc <property>

Safety and privacy notes

The skill runs scripts and uses the network. Commands send requests to Google APIs, and query data or URLs you check can be sent to those APIs. Credentials sit in a plain local config file, so protect that file and the service account key it points to.

Review the scripts before you use the skill on sensitive properties. Do not use it on properties where you do not want query data or URLs sent to Google through scripts you have not reviewed.

  • Scripts run locally and make network calls to Google
  • Keep ~/.config/claude-seo/google-api.json private
  • Check that the scripts folder is present before running commands; the commands are written against a ${CLAUDE_PLUGIN_ROOT} path
  • The skill is MIT licensed

Tips for better results

Start with the credential check so you know which commands will work. If a command fails, missing credentials are the most likely cause.

Read the reference files for details. Search Console data has known gaps: a logging error made impressions, CTR and average position unreliable for a period in 2025 and 2026, so treat trends across that window with caution. For EU sites, the DMA and Consent Mode v2 reference explains how consent changes what Search Console and GA4 record.

  • Run the credential check before any other command
  • Treat Search Console totals as site-wide only when totals_complete is true
  • Report the Lighthouse version from the output instead of assuming the latest
  • Use the Generative AI report for impressions-only AI visibility; you cannot cleanly split AI traffic out of standard totals
  • Read references/rate-limits-quotas.md before running large batches

What people use it for

  • Compare real Chrome field data for key pages against Lighthouse lab scores
  • Track 25-week Core Web Vitals trends after a site release
  • Check whether important URLs are indexed and which canonical Google chose
  • Report monthly organic search clicks and impressions from Search Console for a client
  • Find quick-win queries ranking at positions 4 to 10 with high impressions

Frequently asked questions

Is the seo-google skill free?

The skill is MIT licensed, and the source describes the Google APIs it uses as free. You still need a Google Cloud project and Search Console access to use most commands.

What do I need before I can use seo-google?

You need a Google Cloud project with an API key and/or service account, Search Console access to the property, Python for the bundled scripts, and a config file at ~/.config/claude-seo/google-api.json. GA4 and Keyword Planner commands need extra IDs and tokens.

How do I install the seo-google skill?

Download the ZIP from the directory page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder. In the Claude apps, you can upload it under Settings, then Capabilities, then Skills.

Why do my seo-google commands fail?

Most failures come from missing credentials. Run the credential check first, confirm the config file exists at ~/.config/claude-seo/google-api.json, and confirm your Google account has access to the Search Console property.

Can seo-google replace a crawler like Screaming Frog?

No. The skill adds Google's own data, such as field metrics, indexation status and search performance, to analysis. Crawl-based audits remain useful for finding issues on your own site.

Does seo-google send my data to Google?

Yes, commands send requests to Google APIs, and the query data or URLs you check are part of those requests. Review the scripts first if the property is sensitive.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use seo-google if you already have a Google Cloud project and Search Console access and want real Google field data next to your crawl audits. Skip it if you need a quick one-off check with no setup, or if you are not comfortable storing API credentials locally and sending query data to Google through scripts you have not reviewed.

Download the skill