⌘K

Live evidencecounted 13m ago

1.36days of PURR volume to unwind the crowd99%of the XMR book on one side, across 26 vaults$1.70Bunder one key with no guardian, 103 of the 492 vaults whose control we have read$231.6Mwhose control we have never read, across 3,357 vaults$1.98Bsitting in vaults and not lent out$11.71Bunder 3,849 vaults we read$4.69Bin 335 vaults where we saw automated execution, nothing declared2,238vaults declared or on a venue built to run software, holding $1.07M337software-run vaults scored, 13% of them456operators identified behind them26networks where we read capital, from Ethereum to Cardano32,244,193deposits and withdrawals read from chain184changes recorded in the last day1,655,278ERC-8004 agent registrations read+11.13/ 30 days1,403,733distinct owners behind those registrations204,760registrations carrying a readable card, 12%3,411,376paid requests on the Olas rail (deliveries are counted apart)−0.34/ 30 days88suppliers delivering that work413,151agent-to-agent jobs read on Base−0.08/ 30 days$1.0Msettled between agents, all time1,177suppliers listed on the ACP rail187agents live on Cardano78Cardano agents deregistered, 29% of all ever minted29,885escrows between Cardano agents since Jun 2025, 157 still open75%of every agent payment we read comes from 3,179 addresses that never stop351,892addresses paid once and never came back, of 605,095 that ever paid4,908of those payers are an agent we can name, 0.811% of them312agent hosts answered when we asked, of 2,328 probed17,114agents sit behind a host that answers nothing at all414msmedian answer from the 645 hosts that answer38%of all agent ratings were written by ten addresses, of 808,672 ratings read144,029rated agents have exactly one rater, of 162,234 rated at all130,764service endpoints declared across 56,905 agent cards we read1.36days of PURR volume to unwind the crowd99%of the XMR book on one side, across 26 vaults$1.70Bunder one key with no guardian, 103 of the 492 vaults whose control we have read$231.6Mwhose control we have never read, across 3,357 vaults$1.98Bsitting in vaults and not lent out$11.71Bunder 3,849 vaults we read$4.69Bin 335 vaults where we saw automated execution, nothing declared2,238vaults declared or on a venue built to run software, holding $1.07M337software-run vaults scored, 13% of them456operators identified behind them26networks where we read capital, from Ethereum to Cardano32,244,193deposits and withdrawals read from chain184changes recorded in the last day1,655,278ERC-8004 agent registrations read+11.13/ 30 days1,403,733distinct owners behind those registrations204,760registrations carrying a readable card, 12%3,411,376paid requests on the Olas rail (deliveries are counted apart)−0.34/ 30 days88suppliers delivering that work413,151agent-to-agent jobs read on Base−0.08/ 30 days$1.0Msettled between agents, all time1,177suppliers listed on the ACP rail187agents live on Cardano78Cardano agents deregistered, 29% of all ever minted29,885escrows between Cardano agents since Jun 2025, 157 still open75%of every agent payment we read comes from 3,179 addresses that never stop351,892addresses paid once and never came back, of 605,095 that ever paid4,908of those payers are an agent we can name, 0.811% of them312agent hosts answered when we asked, of 2,328 probed17,114agents sit behind a host that answers nothing at all414msmedian answer from the 645 hosts that answer38%of all agent ratings were written by ten addresses, of 808,672 ratings read144,029rated agents have exactly one rater, of 162,234 rated at all130,764service endpoints declared across 56,905 agent cards we read

Observations

Patterns with more than one explanation. Nothing here lowers a grade.Note 12

The operator holds their own vault

62 vaults · $4.75M

What is measured: The share of the vault held by the address that runs it, at 90% or more. A vault run by another vault is left out: HLP’s sub-vaults are held by HLP itself, which is its depositors’ money and not anybody’s own.
Equally consistent with: A manager with their own money at risk, which is usually a good sign and is why this is listed as context rather than a concern. Not always, though: the address that runs a vault can be a treasury or a custodian holding for others. What the chain shows is who holds the shares, not whose money they are, so this is not by itself skin in the game.
Why it changes a reading: It changes what the track record is evidence of. A return earned almost entirely on the operator’s own capital is a personal account’s performance, not a record of managing anybody else’s money, which is the thing an allocator is trying to judge.

Vault	Platform	Value	What was measured
Deposit. Forget. Wake up richeragent	Hyperliquid	$493.3k	The address that runs this vault holds 100% of it.
PCCagent	Hyperliquid	$51.0k	The address that runs this vault holds 100% of it.
bullmarket tings	Hyperliquid	$44.0k	The address that runs this vault holds 100% of it.
[A] Downsideagent

Showing the 25 furthest past the threshold, of 62.

What this page cannot see

Only vaults whose flow or execution we hold can be examined at all, which is why the denominator above is a few hundred rather than every vault on the site. A pattern that stopped is cleared rather than left standing, so this is a picture of the last thirty days and not a permanent record. And every measure here is a floor: anything held on a centralised exchange, or moved without touching a ledger we read, is invisible to it.