Audit trail
Every verdict the attestation service reaches — charged or refused — is written to a public Hedera topic in order. The topic has no admin key, so nobody can revise or delete what is on it, including us.
- Topic0.0.10451091
- Admin keyNone
- Records shown8
What these records are
Records 25 and 26 are the real history of PLIM-B and PLIM-A: each was read live from the note’s registered vaults on Base. PLIM-B was attested and charged; PLIM-A was refused and not charged, $1.000000 against $1,000,000.00. Records 22 to 24 came before the vault sets were registered and are fixtures — their positions are invented and they say so themselves. Records 17 to 19 are the service’s own test notes. Every record declares its own feed, so nothing here depends on taking our word for which is which.
The register
| Seq | Note | Verdict | Reading | Backing | Charged | What the ledger shows |
|---|---|---|---|---|---|---|
| 25 | PLIM-B | Attested | 140.00% $14.000018 against $10.00 | live, Base block 51,243,979 | Yes, on HashScan | Charged once, for a live reading of the registered vaults: the first attestation of this note’s real backing. |
| 26 | PLIM-A | Refused · coverage_below_floor | 0.00% $1.000000 against $1,000,000.00 | live, Base block 51,244,012 | No | No transfer exists. A live 0.00% that is not an empty wallet: the amounts are the finding. |
| 22 | PLIM-B | Attested | 150.00% $15.00 against $10.00 | fixture:plimsoll-base-backing-v0, retired vault set | Yes, on HashScan | Charged once, and the charge is real. The backing it was charged for is not. |
| 23 | PLIM-A | Refused · coverage_below_floor | 0.00% $15.00 against $1,000,000.00 | fixture:plimsoll-base-backing-v0, retired vault set | No | No transfer exists. The shape of the finding, on fixture backing. |
| 24 | NOTE-INDIA | Refused · source_unavailable | No figure | fixture, no reading | No | No transfer exists, and no figure was ever computed to refuse against. |
| 17 | NOTE-ALPHA | Attested | 130.00% $325,000.00 against $250,000.00 | fixture:plimsoll-vault-flows-v0.3.1 | Yes, on HashScan | The charge is one transfer, and it is on the ledger. |
| 18 | NOTE-BRAVO | Refused · coverage_below_floor | 87.00% $217,500.00 against $250,000.00 | fixture:plimsoll-vault-flows-v0.3.1 | No | No transfer exists. The payment was authorised and never submitted. |
| 19 | NOTE-INDIA | Refused · source_unavailable | No figure | no reading | No | No transfer exists, and no figure was ever computed to refuse against. |
An evidence refusal carries no numeric field at all — not a zero. A reading of zero basis points would say zero-percent coverage, which is indistinguishable from a genuine finding of nothing, so absence is recorded as absence. That is also why a refusal with a figure prints both amounts: record 23 is 0.00%, and it is not an empty wallet.
Check a record without asking us
The verifier needs no key, no account and no access to our infrastructure. It re-reads the anchored evidence from the public mirror node, recomputes the ratio with its own arithmetic, asks the ledger what moved, and prints a verdict.
node dist/bin/verify-charge.js --hcs 0.0.10451091:25 --explain
- Charged and warrantedexit 0
- Refused and not chargedexit 0
- A charge nobody will sign forexit 2Every credit in the window is matched against the records on the topic. A credit no record claims is the discrepancy.
- Nothing to evaluateexit 3Never reported as a verdict.
What one record carries
Each verdict is a single message under the 1,024-byte limit, so it never chunks into pieces nobody reassembles. It holds the whole input set: the positions, the raw readings, the block, the source set, the floor, the ratio, the policy, the full signature, and whether a charge occurred.
- Encodingv3 from record 22The version selects the verification rules, so a record is always checked under the rules it was written with. Version 3 also names the feed and the oracle it reports to, which is how records 25 and 26 declare a live reading and 22 to 24 a fixture.
- Evidence refusalsno numbers
- Earlier recordsSeq 1 to 21Written under the first two encodings and left exactly where they are. The topic cannot be revised, which is the property that makes it worth reading.