agentleFS
superagntsuperagnt/supereo-audit/skills/ai-visibility-audit/SKILL.md

ai-visibility-audit

Skill0 starsChanged 4 days agoMITView on GitHub

Run supereo's free AI visibility audit of a site from your agent, show the person what it found, then take them to their next step with supereo. The audit puts the questions a site's buyers ask to five AI engines (ChatGPT, Gemini, Google AI Mode, AI Overviews and Perplexity) and reports how often each one recommends the brand, who it picks instead and the first fixes, in about a minute, free. Use when someone asks how often ChatGPT or other AI engines recommend their brand, wants an AI visibility, AEO or GEO audit of their site, has just connected supereo, or asks what to do next with supereo.

12 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use local-seo.

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.
  • huiferSkill

    local-seo

    本地 SEO 分析和优化专家。自动检测项目是否需要本地 SEO,分析 NAP(Name, Address, Phone)一致性、本地关键词优化、Google Business Profile (GBP) 优化和本地结构化数据生成。提供本地商家搜索引擎排名优化建议,包括 NAP 标准化、本地关键词策略、GBP 完整性检查、评论策略、地图嵌入和本地 SEO 审计。

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 3 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 12 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run supereo's free audit as the person, show them its results, then walk them to their next step.

    line 11
  2. Very specific:

    Call get_next_step first, with site when they named one.

    line 25
  3. Very specific:

    Before you ask the person for anything, show them what the audit found, from get_audit's results:

    line 42
  4. Distinctive:

    Use supereo's numbers and words as they come, and add nothing they don't hold: no estimates, no rankings, no promises about what a fix will …

    line 51
  5. Very specific:

    Call get_next_step with site.

    line 55
  6. Distinctive:

    Every answer, the pages the engines cite and every fix: the ai-visibility-results skill.

    line 64

“Common” is measured: kept by at least 6 in 10 of the 12 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.

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

AI visibility audit

Tools: get_next_step, start_audit and get_audit, from supereo's MCP server. All three are free, on every plan.

Run supereo's free audit as the person, show them its results, then walk them to their next step.

1. Get the tools

They come from supereo's MCP server, https://supereo.app/api/mcp, signed in as the person. Some clients load MCP tools only when asked: search your tools for start_audit before you decide it isn't there. If it isn't, connect the server:

  • Claude Code: claude mcp add --scope user --transport http supereo https://supereo.app/api/mcp, then /mcp to sign in. The tools load in the next session.
  • Claude (claude.ai, the desktop app, and Claude Code signed in with a Claude account): add it as a connector with this link, then connect and sign in. The desktop app loads it once it's refreshed or reopened.
  • Any other agent: https://supereo.app/install has the steps for each one.

Sign-in happens in the person's browser: supereo emails them a code, and their first sign-in makes their account. There is no API key, so never ask for one.

2. Start where the person is

Call get_next_step first, with site when they named one. Its step says where they are, its say what to tell them, and its do what you do next. Follow do as it's written.

  • start_audit: no audit yet. If they haven't named their site, ask which one: it's the only thing to ask. Then run the audit (step 3).
  • audit_running: an audit of theirs is running, maybe of another site. Follow it with get_audit (step 3, part 2), then do the rest of what do says.
  • Any other step: supereo has audited the site already. When they asked for the audit, get_audit with site reads its latest one: show it as step 4 says, with the day it was checked (checkedAt). On a site supereo tracks, its tracked rates are newer than that audit, and the ai-visibility-results skill reads them. Then go on with the step (step 5).

3. Run the audit

  1. start_audit with site: a domain such as acme.com, or any URL on it.
    • running: it has started. Keep its auditId.
    • in_your_account: the site is a workspace of theirs already, so no new audit ran. get_audit with site reads its latest audit: say the day it was checked. An archived workspace stays archived: get_next_step gives the person the link to restore it.
    • A refusal is one plain line, such as one audit running at a time, another account holding the site, or a limit for the day. Tell the person that line as it is.
  2. get_audit with audit_id, again and again until its status is done or failed, waiting progress.checkAgainInSeconds (about 10 seconds) between calls. The audit takes about a minute, sometimes a few. Now and then tell the person where it is, in the words of progress.line ("Asking five AI engines: 85 of 100 answers in."). Never start it again while it runs.
  3. failed: tell the person failure as it is. start_audit can run it again later.

4. Show the results first

Before you ask the person for anything, show them what the audit found, from get_audit's results:

  • The headline (headline): how often ChatGPT recommended them. Then each engine's count (engines): the answers that recommended them out of the answers judged.
  • What those numbers are, as basis says: one free audit that asked each buyer question once on each engine. They are counts of single answers, so give them as counts ("14 of 17"), never as rates or percentages.
  • Who ChatGPT recommended instead (pickedInstead), most first.
  • The first fix (firstFix: what, why and where), and the fact an engine got wrong (wrongFact) when there is one.
  • The best recommendations (recommendations), and the full report (reportUrl).
  • stillComing: answers still landing. They count once they land.

Use supereo's numbers and words as they come, and add nothing they don't hold: no estimates, no rankings, no promises about what a fix will do.

5. Then the next step

Call get_next_step with site. Tell the person its say, and do what its do says: it names the tool to call or when to give them its link, and some steps have nothing to open yet.

  • pick_plan: while the plans are on sale, plans lists each plan with its monthly price and whether this site can take it, and trial says whether the checkout starts with a free trial. The person picks the plan and enters their card on Stripe's page themselves: give them the link, and never open the checkout or enter payment details for them.
  • done: supereo is set up for the site.

Once they've done a step, call get_next_step again for the one after it.

Where to go from here

  • Every answer, the pages the engines cite and every fix: the ai-visibility-results skill.
  • The buyer questions supereo tracks, and how often buyers search for them: the track-questions skill.
  • Making one of supereo's recommended changes yourself: the ship-recommendation skill.
  • Whether AI crawlers can read a page at all: the aeo-audit skill.

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.