Aeo

Answer-engine optimization (AEO / GEO): make pages that get cited in AI answers (Google AI Overviews, ChatGPT, Perplexity, Claude) not just ranked in blue links. Use when the user mentions "AEO", "GEO", "answer engine optimization", "generative engine optimization", "AI Overviews", "get cited by ChatGPT / Perplexity / Google AI", "AI search visibility", "llms.txt", "answer-first content", "featured snippet", "FAQ schema for AI", or "why isn't my content showing up in AI answers". Every guideline here binds to a named, runnable pseolint `aeo/*` rule, so you can verify it instead of guessing. For building the pages at scale, see the `pseolint` (programmatic-seo) skill; for the audit itself, run the pseolint CLI/MCP.From its SKILL.md

Install

npx -y skills add ouranos-labs/pseolint --skill aeo

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

6 stars6 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.9 KB, as published. Nobody here has run it

Answer-engine optimization, with teeth

Search is splitting in two: classic ranking, and being the source an AI answer cites. AEO (a.k.a. GEO) is optimizing for the second. The May 2026 spam-policy revision put manipulation of AI answers explicitly in scope, so the upside and the risk are both real.

This skill is different from generic AEO advice: every recommendation maps to a pseolint aeo/* rule you can actually run. You don't hope a page is citable; you run pseolint and the rule tells you.

The AEO checklist: each line is an enforced rule

Thresholds live in the engine and drift; cite the rule and let pseolint --explain <rule> (or the MCP rule resource) give the current number.

Lead with the answer. A direct, self-contained answer (≈ first 300 chars, in a <p> right under the heading) is what gets extracted. Enforced by aeo/answer-first.
Be citable, not vague. Concrete, attributable facts: numbers, dates, named sources: are what AI engines quote. Enforced by aeo/citable-facts.
Don't bait, deliver. Headlines that promise an answer the body never gives get skipped (and look spammy). Enforced by aeo/summary-bait.
Modular, extractable structure. Self-contained sections under clear headings extract cleanly; walls of prose don't. Enforced by aeo/content-modularity.
Answer the actual questions. Real FAQ coverage of the query's sub-questions. Enforced by aeo/faq-coverage.
Show freshness. Visible, structured publish/modified dates: AI answers favor current sources. Enforced by aeo/freshness-signals.
Let the answer engines in. Don't block AI crawlers (GPTBot, PerplexityBot, Google-Extended, ClaudeBot) in robots. Enforced by aeo/crawler-access (honors Allow per RFC 9309).
Ship llms.txt. A machine-readable map of your canonical, citable content. Enforced by aeo/llms-txt.

Verify: don't eyeball it

# Crawl the site and get the per-page AEO findings
npx pseolint https://yoursite.com

Or call the pseolint MCP audit tool in an agent and read the aeo/* findings. The same risk gate works in CI via @pseolint/action, so an AEO regression can't ship silently.

The trap

AEO and anti-spam pull in opposite directions if you cheat: stuffing fake FAQs, baiting summaries, or mass-generating "answer" pages is exactly the scaled-content / summary-bait pattern the engine (and SpamBrain) flags. Earn the citation with real, well-structured answers: aeo/summary-bait and the spam/* family are there to catch the shortcut.

Works with

pseolint (programmatic-seo) skill: when you're producing answer pages at scale, design the template there (it covers thin-content, entity-swap, and doorway risk); come back here for the per-page AEO layer.
pseolint engine: the MCP server (@pseolint/mcp), CLI (npx pseolint), and Action (@pseolint/action) are what turn this checklist into a verdict.
OKF rule bundle (/okf) + /methodology: the open knowledge behind each aeo/* rule, with current thresholds.