Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+20 moreClaude CodeCursorCodex CLIMuseDots
Manus
Grok
Grok Bot+17 agentsAI Search Visibility Suite: 0-100 GEO Score + Report
GEO audit + AEO audit in one: score any site 0-100 for AI search visibility in ChatGPT, Perplexity, Claude and Google AI Overviews. Checks llms.txt, AI crawler access, schema, freshness. Client report + fix plan.
$22
AI Search Visibility Suite: 0-100 GEO Score + Report
Example session with this skill installed
geo audit https://plausible.io - 10 key pages, white-label client report with the prioritised fix plan
- Read your context and instructions
- Compiled the ai search visibility
- Generated the data export
Summary
Real audit of plausible.io (privacy-friendly web analytics), run on 17 Oct 2026. 10 pages fetched: homepage, /features, /about, /vs-google-analytics, /privacy-focused-web-analytics, /self-hosted-web-analytics, /blog and three blog posts. All 10 returned HTTP 200. AI Search Visibility Score: 74/100 - GOOD. The base is strong: no AI crawler is blocked, every page has JSON-LD, a canonical, html lang and server-rendered copy (590-1,900 words per page). Points are lost on three fixable gaps: no llms.txt, no Organization entity, and deploy-stamped dates.
AI Search Visibility Score: 74/100 - GOOD
| Category | Score | Weight | Weighted | Status |
|---|---|---|---|---|
| AI crawler access + llms.txt | 80/100 | 20% | 16.0 | STRONG |
| Answerability (AEO structure) | 82/100 | 25% | 20.5 | STRONG |
| Structured data + entity | 60/100 | 20% | 12.0 | FAIR |
| Freshness signals | 48/100 | 15% | 7.2 | POOR |
| Share preview (OG / Twitter) | 79/100 | 10% | 7.9 | GOOD |
| Technical basics | 100/100 | 10% | 10.0 | STRONG |
| TOTAL | 100% | 73.6 -> 74/100 | GOOD |
AI crawler access
| Bot | Status |
|---|---|
| GPTBot, OAI-SearchBot, ChatGPT-User | Allowed (no rule) |
| ClaudeBot, PerplexityBot, Google-Extended | Allowed (no rule) |
| Applebot-Extended, Bytespider, CCBot | Allowed (no rule) |
robots.txt blocks none of the 9 AI crawlers checked and lists the sitemap. /llms.txt returns HTTP 404.
Finding 1 - No llms.txt (P1, about 1 hour)
Evidence: https://plausible.io/llms.txt returns 404.
Fix: publish the starter the suite generated from the audited URLs, after replacing each TODO with a one-line, human-written summary:
# Plausible Analytics
> Starter llms.txt for Plausible Analytics. Edit descriptions so AI crawlers and agents can navigate key pages.
## Key pages
- [plausible.io](https://plausible.io/): TODO - one-line summary for AI agents
- [Features](https://plausible.io/features): TODO - one-line summary for AI agents
- [Vs Google Analytics](https://plausible.io/vs-google-analytics): TODO - one-line summary for AI agents
- [Privacy Focused Web Analytics](https://plausible.io/privacy-focused-web-analytics): TODO - one-line summary for AI agents
- [Self Hosted Web Analytics](https://plausible.io/self-hosted-web-analytics): TODO - one-line summary for AI agents
Finding 2 - No Organization entity anywhere (P1)
Evidence: JSON-LD on 10/10 pages, but the only types are WebSite, WebPage, BlogPosting and Person. 0 Organization nodes, 0 sameAs links, 0 BreadcrumbList. The 4 product/landing pages are marked up as BlogPosting with no author or publisher.
Impact: nothing in the markup ties the brand to its official profiles, so AI engines must guess which "Plausible" a page belongs to.
Fix: add Organization JSON-LD to the homepage (name, url, logo, sameAs, contactPoint), using only profiles and facts you can verify. Paste-ready template: templates/organization.jsonld.
Finding 3 - Deploy-stamped dates (P2)
Evidence: /features, /vs-google-analytics, /privacy-focused-web-analytics and /self-hosted-web-analytics all report the identical dateModified 2026-10-13 09:34:48 UTC, which matches a site deploy rather than four separate edits. The homepage, /about and /blog have no date at all and no sitemap lastmod.
Fix: set dateModified only when the page content really changes, and add a visible "Updated" date + JSON-LD dateModified + sitemap lastmod to the 3 undated pages.
Note: the three 2018-2019 blog posts are correctly flagged STALE, but they are dated founder essays. Leave them as they are.
Prioritised fix plan
| # | Priority | Fix | How |
|---|---|---|---|
| 1 | P1 | Publish /llms.txt | Use the generated starter, hand-write the summaries |
| 2 | P1 | Add Organization JSON-LD to the homepage | name, url, logo, sameAs, contactPoint - verified facts only |
| 3 | P2 | Add short FAQ blocks to key commercial pages | 3-5 real buyer questions, 40-80 word answers |
| 4 | P2 | Add author or organisation attribution | 7 of 10 pages have no byline or Organization/Person schema |
| 5 | P2 | Write answer-first passages under headings | /features has 1 extractable passage, one blog post has 0 |
| 6 | P2 | Stop site-wide date stamping | 4 pages share dateModified 2026-10-13 |
| 7 | P2 | Date the 3 undated pages | Visible date + dateModified + lastmod on the same day |
| 8 | P3 | Fix 5 meta descriptions outside 50-160 chars | e.g. /vs-google-analytics is 173 chars |
| 9 | P3 | Add BreadcrumbList on inner pages | Mirror the URL hierarchy |
| 10 | P3 | Trim titles that truncate when shared | Homepage <title> is 75 chars (target 60 or less) |
Files written to runs/plausible-2026-10-17/out/: report.md (white-label client report), score.json, fixes.csv, crawler-access.md, aeo.md, schema.md, freshness.md, share.md.
This is a readiness audit of public HTML on one date. It does not measure or promise citations in ChatGPT, Perplexity, Gemini or Google AI Overviews, traffic or rankings.
Next steps
- Run
geo llmstxtto finish the llms.txt starter - Run
geo schemafor the Organization + BreadcrumbList templates - Re-run in 30 days with
geo compareto show the client how many points they gained
fixes.csv
CSV · data export
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Agencies sell AI visibility audits for $249 to $3,500 (published prices, Oct 2026: BizWhiz $249, Salienz $800, Profitec $1,000 to $2,100, DerivateX $3,500). This suite runs the technical-readiness part of that audit on any live URL (AI crawler access, llms.txt, schema + entity, answerability, freshness, share preview) and writes a white-label client report, for $22. Unlike free GEO checkers, every point in the score is traced to evidence from your own pages and you get a P1/P2/P3 fix plan. It does not run a buyer-prompt citation panel and makes no citation, traffic or ranking promises.
GEO audit + AEO audit in one: score any site 0-100 for AI search visibility in ChatGPT, Perplexity, Claude and Google AI Overviews. Checks llms.txt, AI crawler access, schema, freshness. Client report + fix plan.
Sample output - real run on plausible.io, 17 Oct 2026 (10 pages)
AI Search Visibility Score: 74/100 - GOOD
| Category | Score | Weight | Weighted |
|---|---|---|---|
| AI crawler access + llms.txt | 80 | 20% | 16.0 |
| Answerability (AEO structure) | 82 | 25% | 20.5 |
| Structured data + entity | 60 | 20% | 12.0 |
| Freshness signals | 48 | 15% | 7.2 |
| Share preview (OG / Twitter) | 79 | 10% | 7.9 |
| Technical basics | 100 | 10% | 10.0 |
Top fixes from that report:
- P1 - Publish /llms.txt (currently HTTP 404)
- P1 - Add an Organization JSON-LD entity with sameAs (none on any page)
- P2 - Stop site-wide date stamping: 4 landing pages share one identical dateModified (2026-10-13)
- P2 - Add answer-first passages + short FAQ blocks on commercial pages
The full report, score.json, fixes.csv and every module file from this run ship inside the pack.
Trust Works in Claude Code, Cursor and Codex · v1.0.0 · October 2026 · tested on a real public site (plausible.io) and an offline demo · stdlib Python only, no API keys · fetches only the pages you list (1 request/sec) · never invents schema, dates or data · no citation, traffic or ranking promises · questions answered within 24h · Cygnix Labs (7 live SEO actors on Apify: apify.com/cygnix_labs)
Your site is built for Google. AI answers read it differently.
ChatGPT, Perplexity, Gemini and Google AI Overviews can only cite pages they can fetch, parse and trust. Most sites lose points in the boring places: a robots.txt rule that blocks an AI crawler, no Organization entity, landing pages with no extractable answer, dates that change on every deploy. This suite checks all of it in one run and turns it into a score and a fix list.
How it works
Give your agent a URL. It fetches the homepage plus the pages you choose (or the first N from your sitemap), runs six modules and returns:
- A 0-100 AI Search Visibility Score with a per-category breakdown and status band
- A prioritised fix plan (P1/P2/P3) with the "how" for each fix, also as fixes.csv for your dev team
- A white-label client report (report.md, your agency name and footer from one JSON file)
- Detail files for every module, so every point in the score can be traced to evidence
Six modules, one skill
- AI crawler access + llms.txt (20%) - robots.txt rules for GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended and more; llms.txt check + starter generator
- Answerability / AEO (25%) - one H1, question-style headings, extractable 40-170 word answer passages, lists/tables, server-rendered text, author/organisation attribution, FAQ blocks
- Structured data + entity (20%) - JSON-LD audit against Google's documented properties, Organization / sameAs / BreadcrumbList gaps, paste-ready templates
- Freshness (15%) - visible dates vs JSON-LD dateModified vs sitemap lastmod, plus detection of fake "build stamp" dates
- Share preview (10%) - Open Graph + Twitter Card QA so links unfurl cleanly when your content gets shared
- Technical basics (10%) - canonical, html lang, viewport, indexable, title
Commands
geo audit <url> · geo quick <url> · geo crawlers · geo llmstxt · geo answerability · geo schema · geo freshness · geo share · geo report · geo compare (month-over-month progress report: score change, fixes closed, fixes still open)
Built for agencies selling GEO / AEO audits
Run the audit before the sales call, send the white-label report, re-run monthly and show the client exactly how many points they gained. Works offline on saved HTML too, for sites behind a login or a staging server.
Who it's for
SEO agencies and freelancers adding AI search audits to their offer · in-house marketing teams checking whether the site is readable by AI engines · developers who want a repeatable, scriptable check before launch.
What it won't do
It does not query ChatGPT or Perplexity for you, estimate "AI traffic", or project a future score. It measures readiness on the date you run it and tells you what to fix first.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed with 3 warnings 7 days ago
Requires
Llmstxt
Yoursite
Python