Open data · version 2026-10-19 · regenerated daily

The daily archive, as files

Everything on this site is computed every morning from the same archive. This page publishes that archive: the rating of every x402 service for every day since 2026-08-16, the daily Agent Economy Index, a daily census of the x402 discovery catalog and the rating of every agent skill in the ClawHub registry. Same numbers as the pages, same day, with a schema and a citation. Method: methodology.

Files

file	what	rows	size
`ratings-daily.csv.gz`	Rating of every service, every day since the first run	96,233	2.1 MB
`aei-daily.csv`	Agent Economy Index, one row per day	30	1 KB
`catalog-daily.csv`	Daily census of the x402 discovery catalog	79	3 KB
`skills-latest.csv.gz`	Rating of every agent skill, as of 2026-10-19	80,547	1.6 MB

datapackage.json describes every column (Frictionless Data Package); CITATION.cff is the citation for reference managers.

Licence

Research, journalism and non-commercial use: free, under CC BY 4.0, with the citation below. Reuse the figures, redistribute the files, build on them; say where they came from and when.

Commercial use: licensed. An organization that uses the dataset in a product, a service or an internal process licenses it per organization, per year, at a fixed price published on this page by 15 November 2026. No negotiation, no sales call, no price by e-mail. The licence funds the observatory and never touches a rating: the rated party never pays, and nobody can buy a tier (policy).

Citation

Fonseca, A. (2026). Agent Economy Report: daily trust ratings and settlement of the agent economy [dataset]. Agent Economy Report. https://agenteconomy.report/s/data/

@misc{agenteconomyreport2026,
  author = {Fonseca, A.},
  title = {Agent Economy Report: daily trust ratings and settlement of the agent economy},
  year = {2026},
  howpublished = {\url{https://agenteconomy.report/s/data/}},
  note = {Dataset, version 2026-10-19, CC BY 4.0}
}

DOI: pending (the Zenodo deposit of this dataset is being prepared; until it exists, cite the URL). For a single page, cite the page URL with its as_of date and the method version shown on the methodology page.

Columns

ratings-daily.csv.gz

day date: Rating day (UTC morning run).
host string: Service host as listed in the x402 catalogs, or the payTo wallet for unlisted entries.
kind string: service (host) or unlisted_wallet.
tier string: AAA to D; BBB is the trust line.
score number: Score 0 to 100 (40 adoption, 25 settlement, 25 centrality, 10 uptime).
outlook string: positive, negative, stable or new (trend of organic agents over 7 days).
flag string: CAPTIVE, SYBIL, NEW, RETIRED or empty.
uptime number: Share of daily probes answered 402 or 200 in the 14-day window (0 to 1).
organic_payers integer: Distinct organic paying wallets in the 14-day window (scanners excluded).
payers integer: Distinct paying wallets in the window, scanners included.
organic_payers_7d integer: Organic payers in the last 7 days of the window.
organic_payers_7d_prev integer: Organic payers in the 7 days before that.
settled_usd_14d number: USDC settled to the entry in 14 days at agent scale (a payment counts at most US$ 1).
centrality number: PageRank-style centrality in the organic payment graph.
retired_resources integer: Listed resources that answered 410 with a past Sunset date.
family integer: Family id when several hosts share the payTo wallet (D-0050); empty otherwise.

aei-daily.csv

day date: Day (UTC).
usd number: USDC settled to x402 services at agent scale (a payment counts at most US$ 1).
gross number: USDC settled before the agent-scale cap (gross flow).
tx integer: Number of settlement transfers counted.

catalog-daily.csv

day date: Snapshot day.
resources integer: Resources listed in the CDP discovery catalog that day.
hosts integer: Distinct hosts among those resources.
paid_calls_30d integer: Sum of the 30-day paid-call counts the catalog reports per resource (gross, as declared).
payers_30d integer: Sum of the 30-day distinct-payer counts the catalog reports per resource.

skills-latest.csv.gz

day date: Rating day.
slug string: Skill slug in the ClawHub registry.
tier string: AAA to D.
outlook string: up, down, flat or NEW.
flags string: Pipe-separated flags (SUSPICIOUS, PIPE_SHELL, OBFUSCATED, UNKNOWN_PAYEE, KEYS, NEW).
score number: Score 0 to 100 (45 adoption, 25 maturity, 20 inherited payment trust, 10 instructions).
downloads integer: Registry downloads, cumulative.
downloads_7d integer: Downloads in the last 7 days (empty when the 7-day reference is missing).
installs integer: Registry installs, cumulative.
stars integer: Registry stars.
versions integer: Published versions.
age_days integer: Days since first publication.
pays_x402 boolean: The skill's instructions pay at least one x402 service.
owner string: Registry handle of the author.

What is not here

The raw payment graph (payer wallet, payTo wallet, day, amount) and the raw liveness probes are not published as files: they identify wallets at a granularity the pages do not show. They are available to researchers on request, under the same licence, with citation. Write to contact@agenteconomy.report.