← Back to the report
Agent Readiness

Gymshark

gymshark.com · E-commerce & Retail

70
Readiness /100
26
Checks passed
3
Checks failed
78
Checks run

Discoverability

Can agents find your site and its capabilities?

failWebMCP / agent-protocol tools discoveredAgent manifests

The page registers no tools via the WebMCP browser API (navigator.modelContext).

Fix: Expose your actions to AI agents with the WebMCP browser API: register tools via navigator.modelContext.registerTool. Optionally also publish a /.well-known/webmcp.json tool catalog as a discovery signpost.

warn/llms.txt has a summary blockquoteAgent manifests

Per llmstxt.org, follow the H1 with a "> short summary" blockquote.

Fix: Add a "> One-sentence description of your site." blockquote under the H1.

passAGENTS.md operating guide for agentsAgent manifests

Found a Markdown guide at /AGENTS.md.

pass/llms.txt guide for AI agentsAgent manifests

Found https://gymshark.com/llms.txt.

pass/llms.txt lists key pages as markdown linksAgent manifests

Found at least one "[name](url)" markdown link.

pass/llms.txt opens with an H1 titleAgent manifests

First line is a top-level "# Title" heading.

passKey pages discoveredKey pages crawl

Crawled 2 of 12 discovered pages beyond the homepage.

infoWebsite platformplatform

Built on Shopify.

Understanding

Can agents understand your content?

fail<html lang> attributeStructured data & HTML

No lang attribute on <html>.

Fix: Add lang to the <html> root (e.g. lang="en") so screen readers and AI translation engines pick the right voice / model.

failSingle H1 on homepageStructured data & HTML

No <h1> tag found on the homepage.

Fix: Add exactly one <h1> that names the page. AI summarisers use it as the document title.

warnHomepage content is server-renderedCrawlability (Google)

Homepage server response contains only 1 words of visible text. AI agents that don't execute JavaScript may see an empty page.

Fix: Consider pre-rendering or server-rendering the homepage so AI crawlers (which typically don't run JS) can see the content. Frameworks: Next.js (Server Components), Nuxt SSR, Astro, or Vite SSG.

warnInternal links on homepageStructured data & HTML

Only 0 internal links on the homepage.

Fix: Add at least 5-10 internal links pointing to your top product, pricing, docs, blog and about pages so an AI agent can navigate to them from the homepage.

warnSemantic landmark elementsStructured data & HTML

Missing landmark elements: footer, nav.

warnKey pages have <title> and <h1>Key pages crawl

1 of 2 pages are missing <title> and/or <h1>: /products.

Fix: Make sure every page sets a unique <title> and exactly one <h1>. The title is how an AI agent confirms it landed on the right page after navigating.

warnKey pages are server-renderedKey pages crawl

1 of 2 pages return <80 words of visible text — agents without JavaScript see an empty page: /products (1w).

Fix: Pre-render or server-render these pages so AI crawlers (which usually don't execute JS) can read them. Frameworks: Next.js Server Components, Nuxt SSR, Astro, or build-time prerendering.

passMarkdown content negotiationOpen agent protocols

Homepage returns text/markdown when requested.

passDOM size is reasonableStructured data & HTML

Homepage contains ~97 elements.

passImages have alt textStructured data & HTML

1 of 1 <img> tags have alt attributes.

passImages have explicit dimensionsStructured data & HTML

1/1 images set width and height — no layout shift.

Operability

Can agents actually operate the page?

failARIA input fields do not have accessible namesAccessibility

Lighthouse flagged this audit — agents may be unable to perceive or operate the affected elements.

Fix: Give ARIA input fields an accessible name.

fail`[role]`s are not contained by their required parent elementAccessibility

Lighthouse flagged this audit — agents may be unable to perceive or operate the affected elements.

Fix: Ensure ARIA child roles are nested inside their required parent role.

failWebMCP tools registered with the browserWebMCP tools

The page registers no tools via the navigator.modelContext browser API, so an AI agent has no structured way to operate it.

Fix: Expose your site’s actions to AI agents with the WebMCP browser API: call navigator.modelContext.registerTool({ name, description, inputSchema, execute }) from your page so an agent can invoke them. Implement it directly, or with a library like the @mcp-b polyfill (https://mcp-b.ai). Spec: https://github.com/webmachinelearning/webmcp.

warnAccessibility tree is agent-navigableAccessibility

2 of 14 agent-critical accessibility audits failed (ARIA input fields do not have accessible names, `[role]`s are not contained by their required parent element) — an agent may be unable to identify or operate the affected elements.

Fix: Fix the failing agent-accessibility audits listed below (accessible names on controls, valid ARIA roles/relationships, nothing interactive hidden from the tree).

warnOverall accessibility scoreAccessibility

Lighthouse accessibility score: 83/100. Agents read the page through its accessibility tree, so this is a proxy for how navigable your site is to an AI agent.

Fix: Resolve the failing accessibility audits below — each one removes an element or relationship an agent would otherwise be blind to.

warnTime to First BytePerformance

833ms (field data from CrUX) — needs improvement.

Fix: Every request an AI agent makes pays this server latency, and agents make many requests per task. Reduce TTFB with edge caching, a CDN, and shipping less server-side work per request.

pass`[aria-*]` attributes match their rolesAccessibility

`[aria-*]` attributes match their roles passed.

passUses ARIA roles only on compatible elementsAccessibility

Uses ARIA roles only on compatible elements passed.

pass`[aria-hidden="true"]` is not present on the document `<body>`Accessibility

`[aria-hidden="true"]` is not present on the document `<body>` passed.

pass`[aria-hidden="true"]` elements do not contain focusable descendentsAccessibility

`[aria-hidden="true"]` elements do not contain focusable descendents passed.

pass`[role]`s have all required `[aria-*]` attributesAccessibility

`[role]`s have all required `[aria-*]` attributes passed.

passElements with an ARIA `[role]` that require children to contain a specific `[role]` have all required children.Accessibility

Elements with an ARIA `[role]` that require children to contain a specific `[role]` have all required children. passed.

pass`[role]` values are validAccessibility

`[role]` values are valid passed.

pass`[aria-*]` attributes are valid and not misspelledAccessibility

`[aria-*]` attributes are valid and not misspelled passed.

pass`[aria-*]` attributes have valid valuesAccessibility

`[aria-*]` attributes have valid values passed.

passButtons have an accessible nameAccessibility

Buttons have an accessible name passed.

passDocument has a `<title>` elementAccessibility

Document has a `<title>` element passed.

passLinks have a discernible nameAccessibility

Links have a discernible name passed.

passCumulative Layout ShiftPerformance

0.040 (field data from CrUX) — good.

infoTool catalog signpost publishedWebMCP tools

No tool catalog was found at /.well-known/webmcp.json (or /.well-known/webmcp). The catalog is a community convention, not part of the WebMCP standard, so this does not affect WebMCP presence.

Fix: Consider also publishing a tool catalog at /.well-known/webmcp.json: a JSON document with a "spec" of "webmcp/0.1" and a "tools" array, where each tool declares a name and a clear description. It is a community convention (optional, not part of the WebMCP standard) that lets crawlers and agents discover your tools without executing JavaScript.

Trust & Security

Can agents safely transact?

warnWeb Bot Auth signature headerOpen agent protocols

No Web Bot Auth signature headers — sites can't verify agent identity.

Fix: Web Bot Auth (IETF HTTP Message Signatures over Signature / Signature-Input) lets you cryptographically verify which agent is hitting you. Several CDNs offer turn-key support; otherwise skip until vendor support matures.

warnHTTP Strict-Transport-Security headerSecurity & trust

HSTS max-age=7889238 is below six months.

Fix: Raise max-age to at least 15768000 (six months). The HSTS preload list requires 31536000 (one year).

warnReferrer-Policy headerSecurity & trust

No Referrer-Policy header on the homepage.

Fix: Add `Referrer-Policy: strict-origin-when-cross-origin` so outbound links don't leak full URLs (including query strings) to third parties.

passOAuth authorization server metadataOpen agent protocols

/.well-known/oauth-authorization-server returns parseable JSON.

passOAuth protected-resource metadataOpen agent protocols

/.well-known/oauth-protected-resource returns parseable JSON.

passCDN detectedSecurity & trust

cf-ray: a1eb7b1fc825fea4-AMS (Cloudflare); Server: cloudflare

passContent-Security-Policy headerSecurity & trust

CSP header present.

passSite is served over HTTPSSecurity & trust

Homepage scheme is https:.

passhttp:// redirects to https://Security & trust

http://gymshark.com → https://us.checkout.gymshark.com/ (2 hops).

passHomepage shows no error outputSecurity & trust

No stack-trace markers in the first 5KB.

passNo mixed-content referencesSecurity & trust

Homepage HTML references no http:// resources.

passTLS certificate is not near expirySecurity & trust

Valid for 32 more days (Aug 23 07:57:03 2026 GMT, issuer: E7).

passNo tech-stack disclosure in headersSecurity & trust

Server / X-Powered-By headers don't leak product version.

passMIME-sniff protectionSecurity & trust

X-Content-Type-Options: nosniff.

passClick-jacking protectionSecurity & trust

X-Frame-Options: DENY.

Commerce

Can agents shop and check out?

passagents.md states checkout needs human approvalUniversal Commerce

Found the human-approval-for-checkout rule.

passagents.md served as MarkdownUniversal Commerce

Content-Type: text/markdown; charset=utf-8.

passagents.md adds store-specific contextUniversal Commerce

Mentions the store ("gymshark") — looks customised beyond the default.

passagents.md declares the UCP versionUniversal Commerce

References UCP / a UCP version or /.well-known/ucp.

passagents.md exposes no secretsUniversal Commerce

No secret/token or admin-path references found.

passShopify agents.md presentUniversal Commerce

Found at https://gymshark.com/agents.md.

passagents.md points to the MCP endpointUniversal Commerce

References the UCP MCP endpoint (/api/ucp/mcp).

passUCP manifest is parseable JSONUniversal Commerce

Manifest parses as JSON.

passUCP manifest presentUniversal Commerce

Manifest reachable at https://gymshark.com/.well-known/ucp.

passUCP declares supported servicesUniversal Commerce

Services: dev.ucp.shopping.

passUCP declares transportsUniversal Commerce

Transports: mcp, embedded.