Methodology

Everything Verifeth publishes is first-party. We ingest every ERC-8004 identity and reputation event on all 26 networks the standard has reached, from each registry's deploy block onward. On the 4 of them that carry USDC settlement, we also ingest settlements involving known agent wallets. No sampling, no third-party feeds.

Reading is a scan, not a subscription, so every figure here has a far end. Our position on Base is block 51,459,728, reached 18 September 2026. Where a stream is further behind than the others, the figure that rests on it is dated by that stream.

Recognition, not replication

Every catalog entry passes one filter: does it teach detection, or a replication recipe? We publish only recognition. Public patterns stay one generation behind our private signatures, the standard security posture.

We report the smaller, true number

When a headline deflates under scrutiny, we deflate it. Excluding infrastructure and non-agent contract flows removes settlement activity from our own aggregates, and requiring multiple corroborating signals against a population baseline leaves far fewer wallets flagged than naive counting marks. Every filter we add shrinks the number we publish and makes it harder to dispute.

This page carried two figures illustrating that, both gated claims of a signed bundle. That bundle has been withdrawn, and the figures went with it rather than being retyped as literals, which is the same discipline the paragraph above describes. We do not state a single currency total for settlement activity: volumes are tracked per instrument class and are never summed across classes, so no such headline would be derivable.

Licensing

Catalog content is CC BY 4.0.