Priority: P2 — NUT-13 residue-collision guard is dead code in gonuts v0.13.0 (Conduition disclosure): CheckCollidingKeysets never invoked; wallet-level fix belongs in gonuts-tollgate, tracked here
I maintainer di solito rispondono entro 1 giorno
@Amperstrand ci sta già lavorando.
Dal 12/10/2026.
Valutazione
- Difficoltà
- 4/5
- Tempo stimato
- 3-5 giorni
- Idoneità per principianti
- 32/100
Direzione di ricerca
Start by tracing keyset registration and persistence in AddMint, keyset refresh, and restore, then locate CheckCollidingKeysets and the existing SaveKeyset guard. The issue requires checking every registration path and adding tests for collisions, unaffected non-colliding registration, rotation, and partial-state prevention; done means collisions are refused before state is written. The proposed keyset-ID verification also needs validation against the applicable ID variants.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
What (all refs at tag v0.13.0)
The deterministic-derivation implementation is exactly the BIP32 shape the Conduition NUT-13 disclosure (https://conduition.io/code/cashu-disclosure/, 2026-01, bounty-paid) describes as attackable, and the mitigation that already exists in this repo is never invoked:
cashu/nuts/nut13/nut13.go:24,40— the keyset ID is reducedmod (2^31 - 1)tokeyset_id_intnut13.go:51-63— pathm/129372'/0'/keyset_id_int'nut13.go:94-116/:72-75—DeriveSecret=.../counter'/0,DeriveBlindingFactor=.../counter'/1nut13.go:119-143—CheckCollidingKeysets(currentKeysetIds, newMintKeysetIds)implements precisely the disclosure's short-term fix (refuse keysets whose2^31 - 1residues collide; also refuses exact duplicate IDs across mints) — and a grep across the whole module (wallet, mint, cmd, tests) finds zero call sites. Dead code.- No wallet-path verification that a mint-fetched keyset ID actually derives from the mint's published keys (
DeriveKeysetId) — only per-proof DLEQ checks, which do not constrain the ID.
Why it matters
Per the disclosure: a malicious mint can pick a keyset ID whose 31-bit residue collides with a target mint's; a wallet that transacts with both then derives the same secrets and blinding factors for both keysets, and the malicious mint can use the target mint's NUT-09 /restore endpoint as an oracle to harvest blind signatures and construct valid proofs at the target mint. Most NUT-13 wallets were affected; the short-term fix advised for every wallet is exactly the collision guard this repo already has, unwired.
TollGate-side exposure (context from OpenTollGate/tollgate-module-basic-go#703 item 7): mints are operator-curated (accepted_mints + trust allow/blocklists) and the daemon does not auto-transfer between mints, so the no-user-interaction variant does not apply. The realistic vector is an operator running an untrusted mint alongside an honest one. Moderate severity; not a release blocker; the wiring is small.
Proposed fix
- Wire the existing guard into every keyset-registration writer —
AddMint, keyset refresh, and restore: before persisting a new keyset, runCheckCollidingKeysetsagainst all already-registered keysets; on collision, refuse registration with an error naming both mints and both IDs. This follows the same pattern as theSaveKeysetmonotonic-counter guard (the #480/#496 writer-audit class: a guard that exists must be on every writer's path). - Cheap hardening: verify mint-fetched keyset IDs against
DeriveKeysetId(pubkeys)where the ID variant permits, rejecting forged IDs at registration. - Explicitly not proposed now: switching derivation to the disclosure's long-term fix (single HMAC over full keyset ID + counter, no BIP32 path). That changes every derived secret and silently invalidates determinism for existing wallets' counter ranges — it needs a versioned derivation plus a migration design, not a patch. Note also that V2 keyset IDs alone do not fix this:
keysetIdToBigIntfunnels V2 IDs through the same2^31 - 1residue.
Required tests
- Registering a second mint with a residue-colliding keyset ID → refused, error names both mints/IDs.
- Non-colliding multi-mint registration unaffected.
- The collision check runs before any counter or derivation state is written (registration refusal leaves no partial state).
- Keyset rotation/refresh that produces a colliding ID → refused.
- An attacker-chosen ID that does not hash to the mint's published keys → rejected by the verification in fix 2.
References
- Disclosure: https://conduition.io/code/cashu-disclosure/
- Short-term/long-term fix definitions quoted in the disclosure ("Improving security of keyset IDs", "keyset ID v2")
- TollGate tracking: OpenTollGate/tollgate-module-basic-go#703 (fund-safety program, deep-dive item 7), #665 (pre-tag readiness — this audit's report)
- Amperstrand/cashu-audit — divergence database context
- Lingua principale
- Go
- Stelle
- 12
- Fork
- 14
- Merge medio
- 1g 4h
- PR unite (30g)
- 244
Preparare l'ambiente
- Nessun Dockerfile né file Docker Compose
- Nessun modello di pull request
- Leggi la guida per i contributori
Come iniziare
- Leggi tutta la issue e poi la guida ai contributi del progetto.
- Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
- Fai un fork del repository e lavora su un branch.
- Apri una pull request che faccia riferimento al numero della issue.
Altre issue di OpenTollGate/tollgate-module-basic-go
-
Session event emits the customer's MAC address in the p tag — spec-invalid value on every payment responseForse già presa @Amperstrand l’ha presa 4 giorni fa. Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 70/100
OpenTollGate/tollgate-module-basic-go#949 ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 68/100
OpenTollGate/tollgate-module-basic-go#833 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 65/100
OpenTollGate/tollgate-module-basic-go#813 ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 70/100
OpenTollGate/tollgate-module-basic-go#804 ·
I maintainer di solito rispondono entro 1 giorno
-
go-battery needs an ndsctl on PATH: TestPurchaseSessionGuardHoldsThroughTheOutcomeUnknownWindow fails on bare hosts (passes with stub)Forse già presa @Amperstrand l’ha presa 8 giorni fa. Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
OpenTollGate/tollgate-module-basic-go#726 · 2 commenti ·
I maintainer di solito rispondono entro 1 giorno
Tutte le issue di OpenTollGate/tollgate-module-basic-go
Issue simili
-
plugins: symlinked plugin directories are skippedForse già presa @lab1207 l’ha presa oggi. Apertabug good first issue
Difficoltà 2/5 1-3 ore Idoneità per principianti 84/100
eugenioenko/ttt#1029 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 67/100
Edwin-metcalf/MetCode#24 ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 Meno di un'ora Idoneità per principianti 82/100
I maintainer di solito rispondono entro 1 giorno
-
needs-operator
Difficoltà 2/5 1-3 ore Idoneità per principianti 67/100
keyxmakerx/Chronicle#1654 ·
I maintainer di solito rispondono entro 1 giorno
-
[Low] CSI plugin builds staging directories from storage_volume_id without a containment checkApertaarea/security kind/bug
Difficoltà 2/5 1-3 ore Idoneità per principianti 76/100
agent-substrate/substrate#2935 ·
I maintainer di solito rispondono entro 1 giorno