# FlexExt AgentScan — Complete Technical Specification & Remediation Corpus Publisher: FlexExt Engineering (`https://flexext.dev`) Tool URL: `https://aichecker.flexext.dev/` Support & Custom Engineering: `shop@flexext.dev` ## 1. Overview & Purpose More users rely on AI assistants (ChatGPT, Claude, Perplexity, Gemini, Copilot) and autonomous shopping agents to research vendors, compare products, and verify business trustworthiness. Traditional SEO tools only test Googlebot HTML crawling. **FlexExt AgentScan** tests whether modern LLM crawlers and protocol-aware AI agents can discover, parse without JavaScript, disambiguate entities, invoke tools, and read live pricing/stock data on any public domain. ## 2. The 6 Evaluation Pillars (100 Points Total) ### Pillar 1: Can AI Find Your Website? (20 Points) 1. **`robots.txt` & AI Search Crawler Access (4 pts)**: Verifies HTTP 200 `/robots.txt` and checks that `OAI-SearchBot`, `ChatGPT-User`, `GPTBot`, `ClaudeBot`, and `PerplexityBot` are not blocked via `Disallow: /`. 2. **`Content-Signal` Directives (2 pts)**: Checks for `Content-Signal: search=yes, ai-input=yes, ai-train=no` machine-readable AI usage rights. 3. **Canonical XML Sitemap Discovery (4 pts)**: Validates `/sitemap.xml` (`` or ``) or custom `Sitemap:` directive in `robots.txt`. 4. **`/llms.txt` AI Knowledge Index (5 pts)**: Validates structured Markdown index per `llmstxt.org`. 5. **`/llms-full.txt` Deep RAG Corpus (3 pts)**: Validates extended Markdown documentation/catalog corpus. 6. **RFC 8288 HTTP `Link` Discovery Headers (2 pts)**: Checks response headers for `rel="describedby"`, `rel="agent-card"`, `rel="mcp-server-card"`, and `rel="api-catalog"`. ### Pillar 2: Can AI Read Your Content? (20 Points) 1. **Server-Side Content Without JavaScript (7 pts)**: Measures readable text rendered in initial SSR HTML without executing client-side JS bundles. 2. **Markdown Content Negotiation (`Accept: text/markdown`) (6 pts)**: Sends `Accept: text/markdown` and verifies if the server responds with clean Markdown (~80% lower token overhead than raw HTML). 3. **Semantic Heading Hierarchy (3 pts)**: Verifies single `

` and structured `

`–`

` sections. 4. **Core Meta & Canonical Signals (4 pts)**: Verifies ``, `<meta name="description">`, `<link rel="canonical">`, and `<html lang="...">`. ### Pillar 3: Can AI Understand Your Business? (18 Points) 1. **`Schema.org` JSON-LD Presence & Syntax (6 pts)**: Parses every `<script type="application/ld+json">` block and flags syntax errors. 2. **Business Identity Graph (6 pts)**: Verifies `Organization`, `OnlineStore`, or `LocalBusiness` with `sameAs` authority links and `ContactPoint`. 3. **Contextual Schemas (6 pts)**: Verifies `WebSite` (`SearchAction`), `BreadcrumbList`, `FAQPage`, `Product`, or `SoftwareApplication`. ### Pillar 4: Can AI Interact With Your Website? (18 Points) 1. **Model Context Protocol (`/.well-known/mcp/server-card.json`) (5 pts)**: Checks MCP server discovery manifest. 2. **Google Agent-to-Agent (`/.well-known/agent-card.json`) (5 pts)**: Checks A2A agent capability card. 3. **RFC 9727 API Catalog & OpenAPI 3.1 (4 pts)**: Checks `/.well-known/api-catalog` (`application/linkset+json`) and `/openapi.json`. 4. **In-Browser WebMCP & Agent Skills Index (4 pts)**: Checks `navigator.modelContext.registerTool` and `/.well-known/skills/index.json`. ### Pillar 5: Is Your Website Secure & Accessible? (12 Points) 1. **HTTPS TLS Encryption (4 pts)**: Enforces HTTPS canonical URL. 2. **Security Headers & `security.txt` (4 pts)**: Verifies `Strict-Transport-Security` (HSTS), `X-Content-Type-Options: nosniff`, and RFC 9116 `/.well-known/security.txt`. 3. **Unblocked Agent Access & WCAG 2.2 / EAA Image Semantics (4 pts)**: Detects WAF CAPTCHA walls blocking bots and audits `<img>` `alt` attribute coverage for European Accessibility Act (EAA 2025) & multimodal AI readiness. ### Pillar 6: Is Your Online Store / Brand AI-Ready? (12 Points) 1. **E-Commerce Product & Offer Schema / Service Clarity (5 pts)**: Verifies `Product` + `Offer` (`price`, `priceCurrency`, `availability`) on stores or structured `ContactPoint` on corporate sites. 2. **Agentic Commerce Feed (`/.well-known/ai-catalog.json`) (4 pts)**: Verifies machine-readable catalog discovery JSON. 3. **OpenGraph Multimodal Cards (3 pts)**: Verifies `og:title` and `og:image`. ## 3. Native Turnkey Plugins by Platform - **WordPress & WooCommerce**: `FlexExt AI SEO & Agent Readiness Pro for WordPress` (`https://flexext.dev/en-gb/product/wp-ai-seo-agent-readiness-pro`) - **Joomla 3, 4, 5 & 6**: `FlexExt AI SEO & Agent Readiness Pro for Joomla` (`https://flexext.dev/en-gb/product/flexext-ai-seo-agent-readiness-joomla`) and Free Edition (`https://flexext.dev/en-gb/product/ai-seo-agent-readiness-free-joomla`) - **OpenCart 4.x & 3.x**: `Complete SEO, AI Search & Agent Readiness Suite for OpenCart 4 & 3` (`https://flexext.dev/en-gb/product/complete-seo-intelligence-suite-for-opencart`) - **Custom Platforms (Shopify, PrestaShop, Magento, Shopware, Shoper, IdoSell, Next.js)**: Custom implementation & dedicated plugin engineering available via `shop@flexext.dev`.