# AgentShoppable readiness dataset, September 2026

Per-store results of the AgentShoppable scanner (public storefront mode, 39 checks, one test cart
created and cancelled per store). Every storefront is resolved before scanning, so results are for
the exact host a shopper lands on; password-locked and closed stores are excluded, not failed.

- top-brands-2026-09.csv: 462 well-known direct-to-consumer brands attempted, 272 confirmed Shopify and scanned.
- shop-directory-2026-09.csv: 838 merchants collected from Shopify's Shop directory (shop.app) across 13 categories, 830 scanned.
- *-summary.md: aggregate tables, including per-check pass/warn/fail counts and a category breakdown.

Columns: domain, tag (category, directory sample only), platform, klass, passed, total, critical,
failed, warnings, productsNeedingAttention, productsScanned, then the status of key checks
(PASS / WARNING / FAIL / NOT_TESTABLE), then error.

klass values: agent_ready (a real cart was created through the store's agent API), classic (catalog
readable, no working agent cart), headless (no usable agent surface: endpoint absent, redirecting or
malformed), blocked (bot protection refused us), not_open (password-locked or closed), unreachable,
woocommerce, other (not Shopify or WooCommerce).

Method, caveats, corrections and the write-up: https://agentshoppable.com/research
Scanner source: https://github.com/danelas/agentready
License: CC BY 4.0. Cite as "AgentShoppable AI Commerce Readiness Dataset, September 2026".