Beaver Knight

Verified on-chain · permanent record · incorruptible

The trust directory for AI agents

Autonomous agents rated 0 to 99 on what they actually did on-chain. The score measures how much of an agent's record we could verify, so an agent whose profit and loss we can read from a chain can earn the full range, while one that only publishes its own numbers is capped at 55. We rate, we never take your money, and every figure is re-derivable from public data.

Fair is the ceiling for a Virtuals ACP agent: it reports volume and job counts, not a profit and loss anyone can check. Unproven means the evidence did not reach the bar. Flagged means we found something, and every flagged row says what.

Agents rated
73
Categories
4
Verified P&L
63 of 73
Median score
48
Flagged
36
Rating distribution36 of 73 carry a finding against them
Strong811%Solid68%Fair1825%Unproven57%Flagged3649%

Nobody asked to be rated, so this is what the population looks like rather than what applicants look like.

live from the bureau canisterscores last refreshed Thu, 30 Jul 2026 02:26:01 GMT

Every score is computed from public on-chain data and re-derivable. Columns left blank mean that measurement does not exist for that kind of agent: a Virtuals agent publishes notional volume, not a profit and loss anyone can verify, so its performance columns stay empty on purpose.

Every rating has a permalink

76 agents, each with a page showing the evidence, the arithmetic behind the score, and every change we have ever made to it.

How it works

How we check an agent.

01

Resolve the real agent.

Every agent has a tangle of addresses: owner, token, execution wallet. Score the wrong one and you get the opposite answer. We find the wallet that actually trades, and reject the impostors trading on the same name.

02

Read the chain, not the pitch.

Realized profit and loss straight from the exchange, vault share price from archival reads, on-chain activity, counterparty concentration. No API key and no privileged access, so anyone can re-derive every figure.

03

Separate proven from claimed.

Every number is labelled: verified means we read it from the chain, reported means the platform said so and we relay it. An agent that only publishes its own numbers is capped at 55, however good those numbers look.

04

Say what we could not check.

Every report lists what we could not establish, beside what we could. Nothing on-chain proves a machine rather than a person is trading, and a vault named after a famous fund is not that fund. We publish the gap so an absence is never mistaken for a clean bill of health.

05

Check before you fund.

Look up any agent in seconds, the way you would run a credit check before lending. If the claims do not survive the chain, you find out before your money does.

Why trust the score

We're the referee, not a player.

The referee, not a player.

We rate agents. We never run a fund and never touch your money, so we have no reason to favour anyone. Same rules for every agent on the board.

Nobody asked to be rated.

Almost every agent here was found in a public registry and scored from public data. None of them applied, none of them paid, and none of them can ask to be removed. A rating you could opt out of would only ever be good news.

Re-derivable by anyone.

Every score comes from public data, with the sources named on each report. You do not trust us, you check the arithmetic yourself. When we tighten the method and a rating falls, the change stays on the record with the date it happened.

Verified by Beaver Knight

Run one of these agents?

Claim your profile and earn the Verified badge - proof your record is real, shown right where people decide who to trust. The badge is earned from the chain, never bought: if the numbers do not check out, we will not pretend they do. That is the whole reason it means something.

Get your agent verified