Sample Agent-Readiness Audit Report

webmcp-ai.dev · Agent Readiness Audit · sample

Agent Readiness Audit: alpinesupply.example

Alpine Supply Co. sells technical outdoor gear — climbing, ski touring, and alpine equipment — through a WooCommerce storefront with roughly 1,400 SKUs and store pickup in two locations.

47/100

Partially ready

18 pages crawled · 21 findings · 5 recommended WebMCP tools

Executive summary

Alpine Supply is discoverable but not operable. Search engines and AI crawlers can find and read most pages (robots.txt and sitemap are healthy), but the three actions that produce revenue — product search, stock checks, and add-to-cart — are invisible to agents: no WebMCP tools are registered, product pages lack Product structured data with price and availability, and the store's React-powered filtering renders client-side, so agents see an unfiltered catalog dump. An agent asked to "find crampons in stock under $200 and add them to my cart" cannot complete any step of that task today. The fixes are well-bounded: three days of work registers the five tools below, completes product schema, and publishes llms.txt — moving the store from 47 to an estimated 85+ and, more importantly, from readable to transactable.

What this costs you today

Shopping agents comparing gear retailers skip stores whose stock and prices they cannot verify programmatically.
"Is this in stock near me?" — your top pre-purchase question — is unanswerable by agents despite the data existing in WooCommerce.
Client-side filtering means agent queries land on your unfiltered catalog, burying your 60 highest-margin products.
Competitors registering WebMCP tools during the Chrome origin trial are collecting the early-mover citations and agent traffic.

Score breakdown

Category	Weight	Score	Checks
Agent capabilities	30%	15/100	1/5 passed
Discoverability	20%	85/100	3/4 passed
Content accessibility	20%	55/100	2/4 passed
Machine comprehension	20%	40/100	2/4 passed
Trust & security	10%	75/100	2/3 passed

Your WebMCP tool blueprint

The tools alpinesupply.example should register, in priority order. Each is ready to hand to your developer or coding agent.

1. search_products critical

Search the catalog with filters agents actually use: query, category, max price, in-stock only. Wraps the WooCommerce Store API — your React filter UI stays untouched.

document.modelContext.registerTool({
  name: "search_products",
  description: "Search Alpine Supply's catalog of climbing, ski and alpine gear. Supports text query, category, price ceiling and in-stock filtering. Returns name, price, availability and URL.",
  inputSchema: {
    type: "object",
    properties: {
      query:    { type: "string" },
      category: { type: "string", enum: ["climbing","ski-touring","apparel","footwear"] },
      maxPrice: { type: "number" },
      inStock:  { type: "boolean", default: true }
    },
    required: ["query"]
  },
  annotations: { readOnlyHint: true },
  async execute({ query, category, maxPrice, inStock = true }) {
    const params = new URLSearchParams({ search: query, per_page: "8" });
    if (category) params.set("category", category);
    if (inStock) params.set("stock_status", "instock");
    const r = await fetch(`/wp-json/wc/store/v1/products?${params}`);
    let items = await r.json();
    if (maxPrice) items = items.filter(p => p.prices.price / 100 <= maxPrice);
    return items.map(p => ({ name: p.name, price: p.prices.price / 100,
      inStock: p.is_in_stock, url: p.permalink }));
  }
});

2. add_to_cart critical

Adds a variant to the visitor's real session cart via POST /wp-json/wc/store/v1/cart/add-item; returns cart totals and the checkout URL. The visitor completes payment — the agent never touches checkout.

3. check_store_stock high

Per-location availability for your two pickup stores — the data already lives in your stock-location plugin; this exposes it as the answer to your most common pre-purchase question.

4. get_shipping_estimate high

Rates and delivery windows by postcode, read-only, wrapping your existing shipping calculator endpoint.

5. check_return_policy medium

Returns your policy terms by product category (climbing hardware differs from apparel) so agents can verify before recommending a purchase.

Findings & fixes (excerpt — 6 of 21)

Impact	Finding	Fix
critical	No WebMCP tools registered on any page	Register the 5-tool blueprint above; start with search_products.
critical	Product pages missing Product JSON-LD (price/availability)	Enable full Product schema in your SEO plugin; verify offers.price and availability render on all 1,400 SKUs.
high	Catalog filtering is client-side only	Expose filtered results as crawlable URLs (?category=…&instock=1) or rely on the search_products tool for agent traffic.
high	No llms.txt	Publish /llms.txt mapping your category pages, policies, and store locations (draft included in appendix).
medium	GPTBot blocked by inherited robots.txt template	Remove the blanket block or replace with deliberate per-bot policy.
medium	No Chrome origin-trial token	Register at developer.chrome.com/origintrials and serve the token via wp_head.

Implementation roadmap

Phase 1 — Quick wins (0.5 day)

Fix robots.txt AI-crawler policy
Publish llms.txt
Add origin-trial token

Phase 2 — Machine comprehension (1 day)

Product JSON-LD across all SKUs
LocalBusiness schema for both store locations
Crawlable filtered catalog URLs

Phase 3 — Agent capabilities (1.5 days)

Register the 5-tool blueprint with polyfill fallback
Verify with the Model Context Tool Inspector + Lighthouse agentic-browsing audit
Re-scan: target score 85+

Produced by webmcp-ai.dev — automated crawl with expert-calibrated analysis, aligned with the W3C WebMCP draft and Chrome's WebMCP developer guidance.

Your site. This depth. About an hour.

Get your audit — $79

The sample report.