Grading other people's services means being accountable for our own mistakes. Every change made to a published listing — because the data was wrong, because a service showed us evidence, or because we found something new — is recorded here with the reasoning behind it.
Found something wrong? Send the evidence and we'll check it. A correction that holds up gets published here with the date and the reasoning — including when the change goes against our own earlier judgement. We don't quietly edit grades.
What isn't a correction: routine score movement from a sweep, a service's status changing, or new data on a listing that was already accurate. Those are the system working, not a fix.
The operator confirms that uPay.rip and jiayou.rip are its own products and states they are now labelled OPERATOR-BUILT rather than Sponsor. The shared-ownership hypothesis raised in the original entry is therefore established and the disclosure is accurate.
The entry classified an architectural property as an incident, which our model reserves for dated, observable events. The operator is correct that routing via Monero is not on-chain mixing and leaves no comparable signature. The residual downstream risk is not differential from any other no-KYC swap in the directory, so flagging a single service for it was a scoring bias. Retained as a neutral chain_hopping attribute.
Both legal documents are now published and were verified independently. The release-candidate version string is no longer present. The incident is kept in the record with its resolution date rather than deleted: the absence on 2026-07-27 was factual.
Nexo's logging policy was recorded as "no logs", which is wrong for a regulated lender that performs full KYC and AML screening on every account — the listing understated how much data the service holds and shares. The frozen-funds field was also set to "risk based" rather than "freezes", despite documented withdrawal blocks including a London court case over roughly $126M in halted redemptions. Both fields were corrected against Nexo's own terms and public reporting, which drops the grade to E. The same review extended the freeze rule across all fund-holding categories: a lender that blocks withdrawals fails its users exactly as an exchange does, and there was no reason for lending to be scored more leniently.