geo-audit
Skill0 starsChanged 49 days agoMITView on GitHub
Audit a site for whether AI search engines can reach, read and quote it. Use when someone asks why ChatGPT or Perplexity does not cite their site, wants a GEO or AEO audit, asks about AI crawler access, or is about to start optimising for AI search and needs a baseline.
21 other projects use a version of this.
Adopted by
Most recently taken up by
- hmzainjamil/geo-seo-claudetheir own version, first seen Nov 15, 2026
- alpit-patel/seo-aeo-geo-kittheir own version, first seen Nov 8, 2026
- Ghanyte/geo-seo-audittheir own version, first seen Nov 8, 2026
- dougwithseismic/geo-kittheir own version, first seen Nov 5, 2026
- Gallyan/trouvabletheir own version, first seen Nov 1, 2026
- jiguang9/geo-audittheir own version, first seen Oct 25, 2026
Used with
Projects with it also use geo, seo-local and 1 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
ORBWEVASkillgeo
Generative Engine Optimization — optimize content for AI search engines (ChatGPT, Perplexity, Google AI Overviews, Gemini). Use when user wants to rank in AI search, improve AI citability, audit content for LLM visibility, or explicitly invokes /geo commands. Triggers on "GEO", "generative engine", "AI search", "AI overviews", "citability", "Perplexity", "ChatGPT search", or /geo commands.
Also in 7 projects that use this
mangollcSkillseo-local
Local SEO optimization: Google Business Profile post generation, 30-day posting calendars, NAP consistency checks, local citation audits, review response templates, local keywords.
Also in 3 projects that use this
megastepSkillseo-schema
Detect, validate, and generate Schema.org structured data. JSON-LD format preferred. Use when user says "schema", "structured data", "rich results", "JSON-LD", or "markup".
Also in 3 projects that use this
Other versions21 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 21
- zubair-trabzada/geo-seo-claude★ 11ksame name
- vellum-ai/vellum-assistant★ 1.5ksame name
- me2resh/apexyard★ 542same name
- Cognitic-Labs/geoskills★ 30same name
- techhorizonlabs/thl-open★ 16same name
- Ghanyte/geo-seo-audit★ 9same name
- greglas75/zuvo★ 7same name
- jiguang9/geo-audit★ 3same name
- alpit-patel/seo-aeo-geo-kit★ 2same name
- dougwithseismic/geo-kit★ 2same name
11 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository4 other files it gives its agents.
- ai-visibility-checkSkill
- answer-gap-contentSkill
- answer-readinessSkill
- llms-txtSkill
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 20 edited versions we read keep these.
0
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 31
Read the rules properly. Disallow: / under User-agent: * blocks every agent that has no more specific block of its own.
“Common” is measured: kept by at least 6 in 10 of the 20 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
GEO audit
Work out whether the AI engines can reach a site, read it, and find something quotable on it. In that order, because the later checks do not matter if the earlier ones fail.
Run the checks in dependency order
A site that blocks GPTBot does not have a schema problem. Fix the top of this list before reporting anything below it.
1. Crawler access
Fetch https://<domain>/robots.txt and check each of these user agents by name. They are separate crawlers with separate jobs, and blocking one does not imply the others:
| Agent | Operator | What blocking it costs |
GPTBot | OpenAI | Training and, in practice, retrieval coverage |
OAI-SearchBot | OpenAI | ChatGPT search results specifically |
ChatGPT-User | OpenAI | Live fetches when a user asks about your page |
ClaudeBot | Anthropic | Claude's index |
Claude-User | Anthropic | Live fetches during a Claude conversation |
PerplexityBot | Perplexity | Perplexity's index |
Perplexity-User | Perplexity | Live fetches |
Google-Extended | Gemini and AI Overviews grounding | |
Bingbot | Microsoft | Bing and Copilot |
Applebot-Extended | Apple | Apple Intelligence |
Read the rules properly. Disallow: / under User-agent: * blocks every agent that has no more specific block of its own. A more specific group wins outright — an agent named in its own group ignores the wildcard group entirely, it does not merge with it.
Also check for the mistakes that look fine:
- A
Disallow:with no path means allow everything. ADisallow: /blocks everything. One character. Crawl-delayis ignored by most of these agents. It is not a block, so do not report it as one.- A 404 on
/robots.txtallows everything. A 500 is treated as a full block by several crawlers, which is much worse than not having the file. - Cloudflare bot rules, a WAF, or a "verify you are human" interstitial can block a crawler while robots.txt reads permissive. Fetch a real page with each user agent string if you can, rather than trusting the file.
2. The page actually resolves for a crawler
Fetch the homepage and two deep pages with a plain client, no JavaScript. Check:
- Status is 200, not a redirect chain ending somewhere unexpected.
- The main content is in the HTML. If the body is an empty
<div id="root">, the content is client-rendered and several of these agents will see nothing. This is the single most common silent failure. - No
noindexin a meta tag orX-Robots-Tagheader. - The canonical, if present, points where you expect. A canonical pointing at a different page hands the citation to that page.
3. What the engines look for once they can read it
llms.txtat the root. Absent is not fatal and its adoption is genuinely contested, so report it as a small opportunity rather than a fault. See thellms-txtskill for what belongs in one.- Structured data.
OrganizationandWebSiteon the homepage,Articleon posts,FAQPagewhere there are real questions and answers,Productwithofferson product pages. Validate it parses; invalid JSON-LD is worth less than none because it looks done. - Answer structure. Whether a page states its answer near the top, has real headings, and contains anything liftable like a comparison table or an FAQ. See the
answer-readinessskill. - Sitemap at
/sitemap.xml, referenced from robots.txt, returning 200 and listing live URLs.
What to report
Ordered by what blocks what, with the fix beside each finding. Not a score with no instructions attached.
For each finding give: what is wrong, the evidence (the exact line of robots.txt, the URL that 404s), what it costs, and the specific change. "Add User-agent: GPTBot / Allow: / to robots.txt" is a finding. "Improve crawlability" is not.
Say what passed too. A list of only problems reads as a sales pitch and hides whether the important things are already fine.
Two honest caveats to include
- Blocking a crawler is sometimes deliberate. Publishers block
GPTBoton purpose. Ask before calling it a fault. - None of this makes an engine cite you. It makes citation possible. What gets cited is usually somebody else's article about the category, so a clean audit is the floor, not the win. Say so, or the report over-promises.
With the Peak Answer MCP connected
If the audit_site tool is available it runs this whole pass in one call and returns answer readiness, crawler access, robots.txt, sitemap, structured data, headings, meta description, readability and entities together. Use it for the baseline, then read the findings yourself rather than pasting them, because the ordering above is what makes a report actionable.
geo_audit is a different thing: it asks a real model ten real buying questions and takes minutes. It starts a background run and returns immediately, so call it once and collect the result later. Never block an audit on it.
Free, no signup: https://peakanswer.com/tools/ai-visibility-audit
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

