Can agents find your site and its capabilities?
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.
Per llmstxt.org, follow the H1 with a "> short summary" blockquote.
Fix: Add a "> One-sentence description of your site." blockquote under the H1.
Found a Markdown guide at /AGENTS.md.
Found https://sokoglam.com/llms.txt.
Found at least one "[name](url)" markdown link.
First line is a top-level "# Title" heading.
Crawled 5 of 12 discovered pages beyond the homepage.
If your llms.txt file does not follow recommendations, large language models may not be able to understand how you want your website to be crawled or used for training. The [llms.txt](https://llmstxt.org/) file should be a Markdown file containing at least one H1 header.
Lighthouse Agentic Browsing audits ran.
Built on Shopify.
Can agents understand your content?
Homepage contains ~4248 elements — well above the 1500-element budget.
Fix: Trim the homepage DOM: an AI agent has to read all of it before acting, and this page is several times over budget. Long product carousels, hidden modal markup and large tracker scripts are common culprits.
Detected 4 heading level skips (e.g. h2 → h6).
Fix: Tighten the heading order so each level only steps down by one. Skipped levels confuse both screen-readers and AI structural extractors.
132 of 138 images have alt attributes (96%).
Fix: Add alt text to the remaining images. Empty alt="" is fine for purely-decorative graphics.
132/138 images have explicit width+height.
Fix: Add width and height to every <img>. Without them the browser doesn't know the aspect ratio and reflows the page on load (CLS).
Homepage returns text/markdown when requested.
Homepage server response contains 2278 words of visible text — content is reachable without executing JavaScript.
Language declared as "en".
386 internal links found.
All four landmark elements present (main, header, footer, nav).
Exactly one <h1> on the homepage.
All 5 reachable key pages have both a <title> and <h1>.
All 5 pages return ≥80 words of visible text without JavaScript.
Can agents actually operate the page?
Lighthouse flagged this audit — agents may be unable to perceive or operate the affected elements.
Fix: Give every link discernible text (avoid bare icons / "click here") so agents know where it goes.
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.
1 of 14 agent-critical accessibility audits failed (Links do not have a discernible name) — 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).
`[aria-*]` attributes match their roles passed.
Uses ARIA roles only on compatible elements passed.
`[aria-hidden="true"]` is not present on the document `<body>` passed.
`[aria-hidden="true"]` elements do not contain focusable descendents passed.
`[role]`s have all required `[aria-*]` attributes passed.
Elements with an ARIA `[role]` that require children to contain a specific `[role]` have all required children. passed.
`[role]` values are valid passed.
`[aria-*]` attributes are valid and not misspelled passed.
`[aria-*]` attributes have valid values passed.
Buttons have an accessible name passed.
Document has a `<title>` element passed.
ARIA IDs are unique passed.
`<frame>` or `<iframe>` elements have a title passed.
Lighthouse accessibility score: 92/100. Agents read the page through its accessibility tree, so this is a proxy for how navigable your site is to an AI agent.
2 of 2 visible interactive form fields have labels.
All audits passed
0.008
0.000 (field data from CrUX) — good.
396ms (field data from CrUX) — good.
49 forms missing annotations
Lists the [WebMCP tools](http://goo.gle/webmcp-docs) registered at the time of analysis.
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.
Can agents safely transact?
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.
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).
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.
/.well-known/oauth-authorization-server returns parseable JSON.
/.well-known/oauth-protected-resource returns parseable JSON.
cf-ray: a1eb74f5884338c0-LHR (Cloudflare); Server: cloudflare
CSP header present.
Homepage scheme is https:.
http://sokoglam.com → https://sokoglam.com/ (1 hop).
No stack-trace markers in the first 5KB.
Homepage HTML references no http:// resources.
Valid for 89 more days (Oct 19 11:39:03 2026 GMT, issuer: YE2).
Server / X-Powered-By headers don't leak product version.
X-Content-Type-Options: nosniff.
CSP frame-ancestors directive present.
Can agents shop and check out?
Found the human-approval-for-checkout rule.
Content-Type: text/markdown; charset=utf-8.
Mentions the store ("sokoglam") — looks customised beyond the default.
References UCP / a UCP version or /.well-known/ucp.
No secret/token or admin-path references found.
Found at https://sokoglam.com/agents.md.
References the UCP MCP endpoint (/api/ucp/mcp).
Manifest parses as JSON.
Manifest reachable at https://sokoglam.com/.well-known/ucp.
Services: dev.ucp.shopping.
Transports: mcp, embedded.