Research
The agent economy is measured here the way a science measures anything: a design fixed before the data, a result that could have failed, and a record anyone can check.
The Orthogonality of Economic Reputation: A Sealed Two-Axis Measurement on the Live x402 Settlement Graph
Two preregistered measurements on the live settlement graph (1.77M events). First, identity and settlement are near-disjoint populations: of 743 addresses that really earn (over $1,000 from 5+ distinct payers), exactly 1 appears in the identity registry, so an identity check misses 99.87% of real economic actors. Neither identity (ERC-8004) nor security verification (ERC-8126) substitutes for performed settlement. Second, the ranking's collusion resistance was validated by a strong structured adversary: 80 attacks spanning multi-hop laundering, recirculating mesh, decoys, and cohort overlap, at budgets from $3 to $750, could not buy a top-10 position; the best purchasable rank was #72 for $15, and more money made it worse. Honest scope and a prior closed break are reported in full.
Preregistration and results sealed as witness-chain generations G064 to G066. Defense internals are gated; verdicts, attack families, and cost ladder are published.
The State of x402: A Classified Settlement Map of the Agent Payment Economy
x402 is the agent-to-agent payment economy, and most of it is not real. We classify every address in the settlement graph into service, buyer, rail, treasury, or dust. Roughly 7% of observed activity is genuine settlement and roughly 93% is dust; around 100 services have ever cleared $1,000, dozens of buyers have each spent over $100,000, and over $16M has moved through rails. The economy is sparse and early: in the most recent seven-day window fewer than half of all-time services were active (a snapshot, not a measured churn rate). The classification is a behavioral heuristic whose independent accuracy is not yet established; the live classification is served only from the API.
Snapshot dataset: huggingface.co/datasets/crestsystems/state-of-x402. Per-address classification accuracy is being validated and tracked; the live classification is gated.
Counterfeit Verifiability in Autonomous Agent Payments: A Preregistered Study of Why Verification Must Be Performed, Not Displayed
Across six preregistered stages and over 2,600 decisions, a counterfeit counterparty that displays the surface of trust beats an honest, genuinely settlement-backed one 99% of the time. Agents pattern-match the costume of verifiability, not the fact. Performing actual verification reverses it (1% to 81%), and under exact surface mimicry only performed verification recovers the truth. The design was sealed to a public hash chain before any data were collected, and the null result is reported in full.
Preregistration and stage data sealed as witness-chain generations G040 to G045.
Crest Deployment Systems LLC · Andy Salvo · Jameson Ackerman · agentrank.info