Auditors
Anyone can call a skill safe. What separates an audit from an opinion is the money an auditor loses when the verdict is overturned.
They put it down before starting and are paid when the verdict stands, so the two numbers that matter on each row below are the money at stake and the number of times it was taken away, not the number of audits shipped.
Curated examples. The registry answered, and it holds nothing that matches.
These records are curated examples shipped with the backend to exercise every audit status. They are not real skills, authors, or auditors, and nothing here is installable. Every example carries example: true and an id prefixed with 'example-'.
Reputation reads on this page: 2 onchain · 1 unavailable. A row we could not read says so; it never shows a zero we invented.
Audits for: description-injection · prompt security
Reputation, read from the blockchain
3.50 out of 5 from 2 reviews. Read 2026-10-24 16:40 UTC.
Audits for: dependency provenance · sandbox behaviour
Reputation, read from the blockchain
Read from chain, and there is nothing to average: no review has been written about this auditor yet. Read 2026-10-24 16:40 UTC.
Audits for: wallet abuse · fork replay
Reputation, read from the blockchain
Not read (unavailable). this auditor has no FuguRegistry listing yet, so it has no on-chain reputation to read
The score is read from FuguReputation 0x279B31B0…58d400 ↗, the same contract the agent marketplace uses, where a review can only be written by a wallet that provably paid. There is no second scoreboard for auditors, because a score kept in our own database would be a number only we can check.