aeo-visibility-review
Skill57 starsChanged 41 days agoMITView on GitHub
Assess how visible a site is in AI answer engines (ChatGPT, Perplexity, Claude, Grok) using OpenGSC's AEO Tracker data, and recommend how to win uncited questions.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- drops-researchSkill
- gsc-performance-reviewSkill
- keyword-researchSkill
- link-prospectingSkill
- page-optimizationSkill
- seo-productionSkill
- site-triageSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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
- Distinctive: line 42
Do not promise citation outcomes.
- Distinctive: line 50
Call it before any other tool.
- Distinctive: line 55
Before any PAID tool, check the research log: if the same research ran within the last 30 days, reuse that result and say so instead of …
≈ “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.
Files
OpenGSC AEO Visibility Review
Goal
Tell the user where they stand in AI search: which tracked questions cite/mention their site per engine, where they're invisible, and what content or authority signals would most plausibly change that.
Required inputs
- The site (domain). Call
list_sitesif unknown.
OpenGSC MCP tools
get_aeo_visibility: tracked questions with the latest per-engine cited/not-cited state. Engines using live web search (ChatGPT, Perplexity) signal citation; Claude/Grok signal brand mention.get_search_performance(dimension=query): questions people already find the site with in Google — candidates to add to AEO tracking.get_striking_distance: pages close to page 1 often correlate with citability — sources AI search tends to pick up.get_geo_audits: stored GEO audit reports. Where AEO tracks the same questions over time, these are one-off deep audits of a single query — they name which competitors AI search cited and why, which is the "what does a citable source look like here" evidence the scoreboard lacks.fetch_page_content: pull a page that did get cited and read what makes it quotable — direct answers up top, named entities, data points, visible dates.
Workflow
- Pull
get_aeo_visibility. Build a per-engine scoreboard: cited / not cited / never checked. - Classify the losses: invisible everywhere (content gap) vs cited by search-grounded engines only (authority/mention gap) vs mentioned but not cited (formatting/source-quality gap).
- Cross-reference with
get_search_performance: strong Google queries with zero AI citations are the highest-leverage fixes — the content already ranks, it needs to become quotable (clear answers, data points, named entities, updated dates). - Suggest 3–5 new questions worth tracking, phrased the way real users ask assistants.
Output format
- Scoreboard table: question × engine (✓ cited / ✗ not / — unchecked).
- Diagnosis per losing question (one line each).
- Prioritized fixes: content changes first, then authority plays.
- Suggested new questions to track.
Guardrails
- An unchecked engine is "no data", not "not cited".
- Engine coverage differs (citation vs mention detection) — do not compare raw counts across engines as if equivalent.
- Do not promise citation outcomes; frame as probability-raising moves.
Site context — read first, free
get_site_context is this skill's step zero. It returns the site's shared memory — business
overview, current goal, positioning, writing preferences, the competitor list, key pages and
the research log — and grounds every decision below in what the operator already told us.
- Call it before any other tool. If it errors or comes back empty, continue without it; never block the workflow on missing context.
- This skill especially needs:
positioning,competitors. If a needed section is empty, do a minimal inline fill — ask the user (or infer from the site and confirm), write it back withupdate_site_context, then continue. Never front-load a full interview. - Before any PAID tool, check the research log: if the same research ran within the last 30 days, reuse that result and say so instead of re-buying it.
- On finish, write back what is durable — new or corrected key pages, competitors the work
revealed, positioning facts the user stated — and append a research-log entry if the
session spent paid units (
appendResearchLog).
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.

