agentsclimarketplace
Sign in

Ai search optimization

Skill social-media-skills/skills/skills/ai-search-optimization

Use to get a brand and its content CITED and RECOMMENDED by AI answer engines — the GEO (Generative Engine Optimization) / AI-search-visibility skill. Run when the user says "GEO," "get cited by ChatGPT/Perplexity/Google AI," "ChatGPT SEO," "LLM SEO / LLMO," "AI Overviews," "answer engine optimization (AEO)," "will AI recommend my brand," or wants to show up in AI-generated answers, not just the feed or Google links. Reads brand-profile and audience first. AI engines retrieve + fan-out and cite community/social sources heavily (Reddit, YouTube, Wikipedia); platforms disagree; earned media beats product pages; extractable, fresh content drives citation. Covers the four GEO levers (retrievable, earned mentions, extractable content, entity clarity), authentic social plays, and the manual prompt-audit method. Refuses astroturfing; never fabricates "share of voice." Sibling of social-seo (platform + Google search). Judges via the audit + GEO tools + AI referral traffic.From its SKILL.md

Install
npx -y skills add social-media-skills/skills --skill ai-search-optimization

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

  • 22 stars22 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.

What its file declares

Copied from the file, not written here

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

8.0 KB, ~1.8k tokens by cl100k_base, as published. Nobody here has run it

AI Search Optimization (GEO)

Get your brand into the answer when people ask ChatGPT, Perplexity, Google AI, Gemini, or Copilot a question in your space — instead of watching a competitor get named. This is GEO (Generative Engine Optimization; also AEO/LLMO): optimizing to be cited and recommended by AI engines. It supplements search/SEO; it doesn't replace it.

Four truths shape everything:

  1. AI answers are built by retrieval + fan-out. Engines retrieve live from search indexes (ChatGPT via OpenAI's own crawler/index, OAI-SearchBot — historically Bing-seeded; Google feeds AI Overviews/AI Mode) and split your topic into sub-queries — so ranking in search feeds AI citation, and you optimize for a constellation of questions.
  2. AI cites community/social sources most. Reddit, YouTube, Wikipedia, LinkedIn, listicles and review sites dominate citations — the cited pages are usually not your pages. Earned mentions beat product pages.
  3. Platforms disagree. ChatGPT skews Wikipedia, Perplexity skews Reddit, AI Overviews lean on E-E-A-T + the community web. Optimizing for one ≠ all.
  4. Extractable, fresh, well-sourced content gets quoted. Quotations, statistics, citations, Q&A structure, and schema lift citation; stale content gets displaced.

(Full mechanics: references/how-ai-engines-cite.md.)

Step 0 — Read the foundation + the goal

Load brand-profile.md and audience.md (entity clarity + the real questions matter). Identify the queries the user wants to be recommended for and the engines their audience uses.

Step 1 — Run the prompt-audit (always start here)

Ask the user's 10–30 buyer-intent queries (plus fan-out sub-questions) across ChatGPT / Perplexity / Gemini in fresh sessions; document whether the brand appears, how it's described, and which sources are cited. The cited sources are the strategy; the gaps are the content list. This is the honest ground-truth method — see references/audit-and-measurement.md.

Step 2 — Be retrievable (the foundation)

If you can't be found in search, you can't be cited: rank in Google/Bing and in platform search → social-seo (the sibling). Same keyword/question research powers both. And verify AI retrieval crawlers can reach the site — robots.txt and CDN/bot-protection defaults (e.g. Cloudflare) often block OAI-SearchBot / ChatGPT-User / PerplexityBot / Claude's bots unintentionally.

Step 3 — Earn brand mentions across cited sources (the social core)

Where AI looks most — done authentically: valuable Reddit participation in buyer-intent threads; YouTube with brand + keywords in titles/transcripts (a top AI-Overview signal); LinkedIn expertise; Quora; earned "best [X]" listicle and review-site (G2/Trustpilot) inclusion; relationship-driven PR. The goal is a web of mutual verification. See references/the-geo-levers.md.

Step 4 — Make content extractable

So a model can lift a clean claim: lead with a TL;DR answer, question-shaped headings, lists/ tables, quotations + verifiable stats + citations (the research-backed levers), FAQ/Article schema, named author + dates, and keep it fresh (citations decay). (This lever spans your website/blog too — broader than social; pair with social-seo.)

Step 5 — Build entity clarity

Give the model a clean entity to recommend: a consistent one-line description across site/profiles/ listings → brand-profile; Wikipedia/Wikidata if genuinely notable; claimed listings + consistent NAP; a corroborated "the X for Y" position.

Step 6 — Measure (honestly) + the boundary

Re-run the audit monthly (expect a multi-week lag; judge over quarters), optionally add a GEO tracking tool, and watch AI referral traffic (chatgpt/perplexity referrers). Never fabricate a "share of voice" or citation %. No WoopSocial analytics. Sibling boundary: social-seo = found in platform + Google search; this = cited by AI answer engines.

Orchestration map

ai-search-optimization sets the AI-visibility layer; it routes to / pairs with: social-seo (retrieval/search foundation — sibling) · brand-profile (entity) · content-pillars (question clusters) · reels-script / the growth skills (the YouTube/Reddit/LinkedIn content that earns mentions) · viral-reverse-engineering (what gets cited/shared) · scheduling-and-queue (publish).

Quality bar — self-check

  • Did I start with the prompt-audit and let the cited sources drive strategy?
  • Did I apply the four levers (retrievable → earned mentions → extractable → entity), foregrounding the community/social plays?
  • Did I respect that AI cites earned/community sources over product pages, and that platforms differ?
  • Did I keep it authentic (refuse astroturfing/fake reviews) and never fabricate share-of-voice numbers?
  • Did I hand the search/retrieval foundation to social-seo, note GEO spans the web too, and use audit/tools/referral measurement (no WoopSocial analytics)?

Edge cases & pushback

  • "Flood Reddit / buy reviews" → refuse astroturfing; it's detectable, removed, and trust-destroying → authentic participation + earned reviews.
  • "Tell me my AI share of voice %" → can't see inside models; run the audit / a tool; don't invent.
  • "Just optimize my product page" → that's ~3% of it; most citations are earned/community sources.
  • "Optimize for AI search" (one thing) → engines differ (ChatGPT≠Perplexity≠AI Overviews); pick the field.
  • "Is this my TikTok/Google SEO?" → related but distinct → social-seo owns platform/Google search.
  • "Does WoopSocial track this?" → no; measure via audit + GEO tools + referral analytics.
  • AI-generated content dump → AI down-weights low-quality AI content; needs human judgment + sources.

Related skills

  • social-seo — the sibling: platform + Google search (the retrieval foundation AI pulls from).
  • brand-profile — the entity/positioning AI must understand; content-pillars — question clusters.
  • reddit-marketing — the how of credible Reddit participation (the top AI-citation source).
  • reels-script, instagram-growth/tiktok-growth/linkedin-growth — the YouTube/Reddit/LinkedIn content that earns the mentions AI cites.
  • viral-reverse-engineering — what gets shared/cited; scheduling-and-queue — publish.

References

  • references/how-ai-engines-cite.md — RAG + query fan-out, which sources get cited, per-engine differences, freshness/decay.
  • references/the-geo-levers.md — the four levers (retrievable · earned mentions/social plays · extractable · entity), with the research-backed lifts.
  • references/audit-and-measurement.md — the manual prompt-audit method, GEO tools, referral traffic, honesty rules.
  • references/examples.md — a worked audit + Reddit/YouTube/extractability/entity plays + honest scope.

What ships with it: 5 files

21.6 KB alongside SKILL.md

evals/

references/

Gives 0 of the 12 instructions most performance cost skills give in ~1.8k tokens

Counted across 797 of the 1,117 authors here whose files we hold, read 2026-09-06

  • Check for product marketing context firstin 46 of 797, across 20 files
  • Measure before optimizingin 31 of 797, across 25 files
  • Profile first to identify the actual bottleneckin 23 of 797, across 22 files
  • Verify your robots.txt allows AI crawlersin 21 of 797, across 12 files
  • Import directly and avoid barrel filesin 19 of 797, across 15 files
  • Spawn all runs in the same turnin 18 of 797, across 11 files
  • Write a draft of the skillin 17 of 797, across 10 files
  • Understand the user's intentin 17 of 797, across 10 files
  • Use React.cache for per-request deduplicationin 16 of 797, across 11 files
  • Profile before optimizingin 16 of 797, across 14 files
  • Include specific numbers with sourcesin 15 of 797, across 8 files
  • Add lazy loading to below-fold imagesin 15 of 797, across 10 files

Said here and by no other author read

  • Load brand profile and audience files
  • Run the prompt audit in fresh sessions
  • Verify AI retrieval crawlers can reach the site
  • Earn brand mentions across cited sources
  • Make website and social content extractable
  • Build a clean and consistent entity clarity

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.