Unsolicited sample — produced without engagement; not client work · Re-audited 2026-07-23 with the current engine: 19/20, unchanged.
anveshan.farm scored 19/20. Its own returns policy is missing the contact email it names.
Taz ran a public-page-only Store Readiness Audit on anveshan.farm on 2026-07-12, as an unsolicited proof-of-work sample. No login, no write access, no relationship with Anveshan. Read-only crawl of publicly reachable pages only.
What we found
Anveshan's storefront is unusually agent-forward — it publishes an agents.md file and advertises a UCP/MCP endpoint in robots.txt, alongside deep structured data (30+ distinct JSON-LD types including MerchantReturnPolicy, NutritionInformation, ShippingDeliveryTime).
But we found a real, concrete gap in the published returns policy itself:
"The policy is one paragraph (1,098 characters)... the sentence reads 'For exchange/return, you can contact us at .' — the contact email is missing from the published policy. A support bot or shopping agent reading this page cannot tell a customer how to initiate a return."
We also flagged a wording-risk item for the client's own legal/compliance review (not legal advice from us):
"Collections are named 'Diabetic Care', 'Immunity-Boosting', and 'Weight Management'. AI shopping agents will repeat these category names as if they were product health claims. Recommend the client review this wording with their own compliance/legal advisor."
Why it matters
A missing email in a live returns policy isn't a copy-editing nitpick once an AI agent is the one reading it on a customer's behalf — the agent can't infer a blank. And category names that read fine to a human shopper get parroted verbatim by a shopping agent as if they were medical claims. Both are one-line fixes once someone reads the site the way an agent does.