Generate Validated JSON-LD Schema Markup for Rich Results

Schema Generator Skill generates, writes and validates structured data markup in JSON-LD for any page. It activates on requests such as generate schema, create JSON-LD, schema markup for a page, structured data, FAQ schema, product schema, article schema, help Google understand my page, rich results schema, or add schema to a page. It outputs validated JSON-LD script blocks, a validation summary, an issues list, implementation instructions and a testing checklist.

Use it when adding or validating JSON-LD structured data for a page whose visible content supports the selected schema type. Structured data must match visible, truthful page content, and valid JSON-LD does not guarantee rich results, indexing or display. The skill never emits fabricated review, rating or product claims, and any unresolved placeholders are reported as blocking issues.

It treats two types as the AEO priority: FAQPage and HowTo are the schema types most directly responsible for AI engine extraction, so FAQPage should always be included on landing pages and blog posts that contain an FAQ section.

The input is a JSON object with a page type (landing page, blog post, pillar page, product page, homepage, FAQ page or how-to page), a list of schema types, and page data. Page data can include the canonical URL, title, description, primary keyword, published and modified dates in ISO format, author name, brand name, brand URL, logo URL, a product (name, description, image, price, ISO 4217 currency, availability and optional SKU), reviews, an aggregate rating, FAQs, a how-to (name, description, ISO 8601 total time and steps), and breadcrumbs. If the schema types list is empty, the skill recommends types from the page type using the menu.

Supported types and their rich results:

• Article for blog posts, pillar pages and news: article rich result and Top Stories.
• FAQPage for FAQ sections: an FAQ accordion in search results.
• HowTo for step-by-step guides: a step-by-step rich result.
• Product for product or pricing pages: price, availability and rating.
• Review and AggregateRating: star rating, with a review count for aggregates.
• BreadcrumbList for pages with a navigation hierarchy: a breadcrumb path.
• Organization for home or About pages: brand knowledge panel signals.
• WebPage for basic structured identity, and WebSite for the homepage only, unlocking a sitelinks search box.

Eligibility rules list the fields Google requires, and missing required fields disqualify the page. FAQPage needs a main entity of Question objects with question text and an accepted answer, at least 2 question-answer pairs, and recommends answers under 300 words. Article needs a headline under 110 characters, an image, a published date, an author and a publisher organization with a logo, and recommends a modified date. Product needs a name, image, an offer with price and currency, a numeric price and an ISO 4217 currency code, and recommends description, availability and aggregate rating. HowTo needs a name, an array of steps and text for each, and recommends total time and step images. BreadcrumbList needs list items with a position starting at 1, a name and an absolute URL.

The output has a fixed format: a header naming the page, URL, types generated, rich results unlocked and AEO signals added; a validation summary table; an issues section with severity and exact fixes; JSON-LD output with one script block per type; implementation instructions; and a testing checklist.

Implementation means placing each script block in the head of the HTML, one block per type and never combining types, testing with Google's Rich Results Test and the Schema.org validator, and requesting re-indexing in Search Console after deploying. The checklist covers required fields, no placeholder text, absolute URLs, ISO 8601 dates, decimal prices such as 49.00, ISO currency codes, at least two FAQ pairs, and no HTML tags inside JSON-LD string values.

A references folder holds pre-validated templates for ten types: FAQ, article, product, review, aggregate rating, how-to, breadcrumb, organization, webpage and website. Placeholders in double braces are replaced by a schema_builder.py script using the page data. The skill receives output from the landing page writer, blog writer, content quality auditor and internal linking skills and is the final step before a publish checklist.

SEO specialists, developers and content teams who need correct, validated structured data tied to real page content.