Methodology

How readiness is scored

Two layers. The free probe is a public snapshot. Store Guard is the attested grade with a fix list and optional registry receipt. DEMO rows are always labeled DEMO.

Free probe vs Store Guard

• Free probe — multi-surface HTTP checks, letter grade, no paid attestation. Runs at /agentic-commerce.
• Store Guard ($149) — weighted category grades, bilingual fix list, sha256 receipt, registry path when earned.
• Not a merchant UCP host— we attest whether agents can read you. We do not publish your storefront's /.well-known/ucp for you on steadywrk.app.

Probe checks

1. UCP discovery — Looks for /.well-known/ucp (or documented equivalents). Absence is a fail for agent-purchase rails — not proof we operate your merchant endpoint.
2. ACP discovery — Looks for /.well-known/acp.json. Complements UCP in the multi-protocol readiness set.
3. Product schema — Product / Offer structured data readable from the storefront HTML. Agents skip catalogs they cannot parse.
4. Robots / crawl access — Whether answer engines and agent crawlers are blocked from money paths.
5. llms.txt / agents maps — Presence of answer-engine maps (llms.txt, agents.md, related discovery files).
6. Trust signals — security.txt and related public trust surfaces when present.
7. Open Graph — Basic share/identity tags agents and messengers reuse when recommending a store.
8. MCP / OpenAPI discovery — Whether machine-readable tool/API surfaces exist for agent integration.

Compared to diagnose-only tools

Platform checkers that score product pages without attestation are useful diagnostics. STEADYWRK's lane is bilingual MENA-facing grades, forwardable registry receipts, an MCP-callable probe, and a priced Store Guard SKU — not another free pass/fail clone.