SML-4B · Issue 000 · Measurement standard

A payment is not commerce.

Agent commerce is a settlement joined to a delivery receipt.

Issue 000 freezes the counting rule. It does not infer intent from chain volume, does not turn relayer activity into buyer demand, and does not publish an accepted total before SML has indexed and joined its own first measurement window.

The counting rule

Accepted commerce = settlement ∩ delivery receipt. A chain settlement is necessary but not sufficient. The settlement must be joined to evidence that the paid delivery was accepted under the measurement contract. If the join is absent, the row does not move into accepted commerce.

BRIDGE

Value movement identified as bridge or cross-domain transfer activity. It remains bridge activity, not commerce.

FACILITATOR WALLET

Settlement or submission activity attributable to a facilitator or relayer wallet. Infrastructure movement is not promoted to commerce.

DUST CYCLE

Repeated low-value cyclic movement that meets the published dust-cycle rule. It remains a dust cycle unless a delivery receipt proves otherwise.

UNATTRIBUTED

Settlement evidence exists, but the evidence is insufficient to classify it into another bucket or join it to an accepted delivery receipt.

Accepted is deliberately empty in Issue 000

ACCEPTED: —

No accepted total is quoted in Issue 000. The first accepted total belongs to Issue 001, after SML indexes a defined window and joins settlements to delivery receipts using this frozen rule.

Named-wallet census

Wallet classifications are append-only and evidence-bearing. Every wallet entry must name the address, bucket, evidence source, observation window, classifier version, and superseding record if reclassified. Aggregate totals never erase the underlying named-wallet record.

Allowed settlement states: accepted bridge facilitator_wallet dust_cycle unattributed Promotion rule: settlement -> accepted ONLY IF delivery_receipt_join == true No receipt join -> accepted total unchanged.

Why this standard exists

The population-scale x402 study by Shengchen Ling, Yajin Zhou, Lei Wu, and Cong Wang measured 136.7 million settlements worth about $44.1 million over 280 days and reported that 21.2% were fictitious and 63.78% were internal to linked clusters. Those results show why raw settlement counts and values should not be treated as autonomous commerce without a stronger counting rule.

Ling, Zhou, Wu, Wang — “How Agentic Is Agentic Commerce?” arXiv:2607.12575 →

Evidence hold: SML will not publish the proposed 54.8× / $42.3B relayer-filter sensitivity figure in Issue 000 until its reproducible source artifact is linked. The rule does not depend on that number.

Issue sequence

Issue 000: freeze the ontology, join rule, append-only wallet census format, and no-fabrication boundary.

Issue 001: publish the first SML-indexed measurement window, named-wallet classifications, receipt joins, bucket totals, and accepted total. Until that file exists, accepted remains unquoted.

Outreach line

A payment is not commerce. Agent commerce is a settlement joined to a delivery receipt. The other four buckets have names. https://scriptmasterlabs.com/sml-4b

Machine contract

https://scriptmasterlabs.com/sml-4b.json

Truth First. Proof Always. Pay Only for Accepted Delivery.

SCRIPTMASTERLABS · READ · EXPLORE · VERIFY · EARN

Books, free tools, sources & affiliate opportunities