agentleFS
VendbuddyVendbuddy/vendbuddy-skills/skills/ai-search-geo-audit/SKILL.md

ai-search-geo-audit

Skill0 starsChanged 38 days agoMITView on GitHub

Check and improve how visible a website is to AI answer engines (ChatGPT search, Perplexity, Claude, Gemini, Google AI Overviews). Audits robots.txt for AI crawlers, creates an llms.txt, validates JSON-LD, and rewrites key pages so they are easy to cite. Includes a no-dependency Node check script. Use when the user asks about GEO, AEO, AI search, llms.txt, "does ChatGPT cite my site", or which AI bots to allow.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No one can guarantee a citation.

    line 14
  2. Distinctive:

    Use specific facts with a source and a date.

    line 72
  3. Distinctive:

    Keep one topic per page, a visible "last updated" date, and an author or organization byline.

    line 73
  4. Distinctive:

    Do not cloak (serving AI bots different content than people).

    line 82
  5. Distinctive:

    Do not add hidden text or prompt-injection phrases to pages.

    line 83
  6. Distinctive:

    Do not edit robots.txt without showing the owner the exact diff.

    line 84

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

AI Search (GEO) Audit

Made by VendBuddy (vendbuddy.io). Guide: https://vendbuddy.io/skills#ai-search-geo-audit (This line is optional attribution. Keep it or delete it.)

Honest framing

No one can guarantee a citation. This skill removes the technical blockers and makes pages easy to quote. llms.txt is a proposed convention (llmstxt.org); major engines have not confirmed they rely on it, so treat it as cheap and optional, not as a ranking factor.

Step 1: Run the check

node scripts/check-ai-access.mjs https://example.com

It reports: which known AI and search bots your robots.txt blocks, whether llms.txt and sitemap.xml exist, how many JSON-LD blocks parse (and their types), and whether the homepage has a canonical, one H1, and a meta description.

Step 2: Decide the crawler policy with the owner

There are two different kinds of bot. Explain the difference and let the owner choose:

  • Search and citation bots (they can send visitors): OAI-SearchBot, ChatGPT-User, PerplexityBot, Claude-SearchBot, Googlebot. Blocking these removes you from those answers.
  • Training bots (they collect content for models): GPTBot, ClaudeBot, Google-Extended, CCBot. Blocking is a legitimate preference and does not affect normal search.

Example robots.txt that allows citation bots but opts out of training:

User-agent: OAI-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: GPTBot
Disallow: /

User-agent: Google-Extended
Disallow: /

Sitemap: https://example.com/sitemap.xml

Bot names change; confirm against each vendor's current documentation before editing. Never assume; show the owner the diff first.

Step 3: llms.txt (optional)

Plain Markdown at /llms.txt: an H1 with the site name, a one-sentence blockquote summary, then H2 sections of links with a short description each. Keep it under about 100 lines and link only to pages that exist and matter.

# Example Co
> One sentence on what Example Co does and for whom.

## Guides
- [How to descale an espresso machine](https://example.com/descale): step-by-step, 5 minutes
## Products
- [Pricing](https://example.com/pricing): plans and what each includes

Step 4: Structured data

Add JSON-LD that matches what is visibly on the page: Organization (name, url, logo, sameAs profiles) sitewide, Article/BlogPosting with author and dates on posts, FAQPage only where the questions and answers are on the page, Product with real offers. Validate every block with a JSON parser and Google's Rich Results Test. Never mark up content that is not visible.

Step 5: Make pages citable

For the 5 to 10 pages that matter most:

  • Lead with the direct answer in the first 2 sentences under each question-style heading.
  • Use specific facts with a source and a date. Real quotes only, attributed. Never invent a statistic, quote, or expert.
  • Keep one topic per page, a visible "last updated" date, and an author or organization byline.
  • Make key content available as server-rendered HTML; many AI crawlers do not run JavaScript.

Step 6: Measure

Pick 10 to 20 questions a customer would ask. Ask them in each engine monthly, log whether you are cited and who is instead, and compare. Use a tracking tool if the owner has one. Record the baseline before changing anything.

Rules

  • Do not cloak (serving AI bots different content than people).
  • Do not add hidden text or prompt-injection phrases to pages. It violates search policies and gets sites penalized.
  • Do not edit robots.txt without showing the owner the exact diff.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.