Help us improve
Share bugs, ideas, or general feedback.
From marketing
Structure content so AI assistants and answer engines (ChatGPT, Claude, Perplexity, Google AI Overviews, Bing Copilot) can extract, attribute, and cite it cleanly. Covers claim density, question-led structure, schema markup, citation hygiene, and "answer chunks" that survive being lifted out of context. Use this skill any time the user wants content to perform in AI/LLM-driven discovery, even if they say "make this AI-friendly," "rank in ChatGPT," "answer engine optimization," or "GEO." Trigger whenever content is being drafted or restructured for an audience that may arrive via an AI summary instead of a classical search result, so the page is built to be cited rather than just clicked.
npx claudepluginhub bpainter/composable-dxp-claude-marketplace --plugin marketingHow this skill is triggered — by the user, by Claude, or both
Slash command
/marketing:marketing-geo-contentThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
GEO is the discipline of making content that an LLM-driven answer engine can confidently quote. The goal is no longer just "rank on Google" — it is "show up as the cited source inside an AI answer." The content patterns that win are different enough from classical SEO that they need their own skill.
Provides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
Explores codebases via GitNexus: discover repos, query execution flows, trace processes, inspect symbol callers/callees, and review architecture.
Share bugs, ideas, or general feedback.
GEO is the discipline of making content that an LLM-driven answer engine can confidently quote. The goal is no longer just "rank on Google" — it is "show up as the cited source inside an AI answer." The content patterns that win are different enough from classical SEO that they need their own skill.
Pair this with strategy-seo-brief for the keyword and SERP layer. They are complementary, not competing.
An answer engine doesn't read your page top to bottom and decide if it likes you. It pulls chunks — usually a paragraph or a list — and either uses them to compose an answer or quotes them with attribution. Three implications:
Phrase H2s and H3s as questions a real person would ask. Models match queries to question-shaped headings far more readily than to label-shaped ones.
| Don't | Do |
|---|---|
| Vesting basics | What is founder vesting? |
| Equity considerations | How much equity should a co-founder get? |
| Tax implications | When should I file an 83(b) election? |
Each section begins with a 1-2 sentence direct answer, then follows with context. The first sentence is the chunk most likely to be lifted — it should be self-contained.
What is founder vesting?
Founder vesting is a schedule that determines when a founder's equity actually becomes theirs to keep — typically 4 years with a 1-year cliff. It exists so that a co-founder who leaves after six months doesn't walk away with a quarter of the company.
The first sentence answers the question. The second adds the trust-building "why" without requiring it.
Models love numbered or bulleted lists when each item is a complete claim, not a sentence fragment.
The most common founder vesting terms are:
1. Four-year total vesting period.
2. One-year cliff (no vesting until month 12, then 25% vests at once).
3. Monthly vesting after the cliff.
4. Acceleration on certain triggers, often double-trigger on acquisition.
Each line could be lifted alone and still mean something.
Frequency-anchored statements are highly liftable because they read as factual market norms. They also stay safe for any audience that requires general statements rather than prescriptive advice (regulated industries, legal-adjacent content, healthcare).
Most mid-market enterprises adopt composable DXP in stages — starting with a single channel rather than a full replatform. Some lead with content (CMS-first); some lead with commerce. Almost nobody ships best-of-breed across every layer in the first phase.
A bullet block at the top of the page summarizing the 3-5 main claims. This is the single most-cited block on most well-optimized pages.
Two-axis comparisons (entity types, financing instruments, vesting variations) belong in tables. Models extract them cleanly and reproduce them faithfully.
A 4-8 question FAQ section using FAQPage schema. These get pulled into Google's AI Overviews and "People Also Ask" features, and they double as anchors for direct LLM citation.
Aim for content where most paragraphs contain a verifiable claim. Avoid prose that is connective tissue without substance ("In today's startup ecosystem, founders face many choices.").
A useful self-check: if you removed every sentence that doesn't either (a) state a fact, (b) make a claim about what is typical, or (c) define a term — would the page still hold up? If yes, you have good claim density. If half the page disappears, rewrite.
Models cite sources they trust. Build trust via:
For any regulated or legally-adjacent content (legal, healthcare, finance), AI-assisted authoring is prone to hallucinating citations — every citation gets human verification before publish. Never let an AI-drafted citation into production unverified.
Use schema.org JSON-LD on every published page. The combinations that matter most for founder-facing content:
Article (or BlogPosting) — author, datePublished, dateModified, headline, description.FAQPage — for the FAQ section. Each Q&A becomes a Question + Answer node.BreadcrumbList — site hierarchy.HowTo — only when the content is genuinely a step-by-step procedure (e.g., "How to file an 83(b) election").Organization — site-wide, includes legalName, sameAs links.Person — for author entities, with credentials and affiliation.WebPage with speakable annotations — flag the most important sentences for voice/snippet extraction.If working with Contentful, model these as part of the content type so authors don't have to hand-author JSON-LD.
LLMs ingest content from many surfaces. Where reasonable, also produce:
These are scaffolding, not a substitute for content quality, but they remove friction for the engines that want to ingest you.
Run this before publish:
strategy-seo-brief.dev-contentful-model skill.