bring signatory up to date with the remote signer spec #1257

lescuer97 · 2025-11-08T17:07:15Z

Description

Changes the signatory keyset generation and updates the proto spec to follow this pr

1- Generates new keysets with the correct derivations paths.
2- returns the derivation_index as version
3- add test vectors from remote signer

Notes to the reviewers

This should not affect any already created keyset.

Suggested CHANGELOG Updates

CHANGED

ADDED

REMOVED

FIXED

Checklist

I followed the code style guidelines
I ran just final-check before committing

thesimplekid · 2025-11-12T17:11:13Z

crates/cashu/src/nuts/nut00/mod.rs

+    ///  Big endian encoded integer of the first 4 bytes of the sha256 hash of the unit string.
+    pub fn derivation_index(&self) -> u32 {
+        use bitcoin::hashes::sha256;
+
+        let unit_str = self.to_string().to_lowercase();
+        let bytes = <sha256::Hash as BitcoinHash>::hash(unit_str.as_bytes());
+        // Take the first 4 bytes and convert to u32 (big endian) make sure the integer
+        u32::from_be_bytes([bytes[0], bytes[1], bytes[2], bytes[3]]) & !(1 << 31)


Is this spec'd? @vnprc has been working on something similar. #1004

yeah, it has always been part of the remote signer spec: https://github.com/cashubtc/nuts/pull/250/files.

The only new addition is the modulo on the sha256sum. @lollerfirst had a look at that and gave me some thumbs up on it

@lescuer97 This will be a breaking change for the known types, how are planning to address that?

yes I have open PRs to cdk and the NUTs repo to implement deterministic derivation paths
cdk PR #1181
NUTs PR cashubtc/nuts#292

yes I have open PRs to cdk and the NUTs repo to implement deterministic derivation paths cdk PR #1181 NUTs PR cashubtc/nuts#292

I was not aware of this PRs. @vnprc would you like to use mine? they seems to be equally as effective.

@lescuer97 This will be a breaking change for the known types, how are planning to address that?

you are right, it would break it for library users. I could make a new version and make a deprecation warning for the older function. What do you think?

@crodas I added back and depracted the original index function.

thesimplekid · 2025-11-14T16:29:22Z

crates/cashu/src/nuts/nut00/mod.rs

+    #[deprecated(
+        since = "0.14.0",
+        note = "This function is outdated; use `hashed_derivation_index` instead."
+    )]


I don't think we should deprecate this. Instead we keep the hard coded for the ones we already have and then hash anything else.

I don't think that would be a good idea. It would not make cdk spec compliant.

I don't think breaking types here is a big deal thought.

As an alternative, I can also just do a simple hack and return Some(u32). even thought there is no option. This would mantain type consistent

I don't think that would be a good idea. It would not make cdk spec compliant.

There is no merged spec so this is not an issue.

I still have concerns about removing support of the legacy method for known units. Both nutshell and cdk use the same paths now even though it is not spec'd. So it would be possible to keep that in the spec as a legacy path. The known units keep the current method anything else goes this way.

Right now if only the signatory db is lost from before this spec where there are legacy keysets after this change these keysets would not be recoverable and the mint could not sign or verify ecash from them.

thesimplekid · 2025-11-17T11:04:10Z

Are there still open questions on this should we bump it to 0.15?

vnprc · 2025-11-17T19:32:17Z

crates/cashu/src/nuts/nut00/mod.rs

+    ///  Big endian encoded integer of the first 4 bytes of the sha256 hash of the unit string.
+    pub fn hashed_derivation_index(&self) -> u32 {
+        use bitcoin::hashes::sha256;
+
+        let unit_str = self.to_string().to_lowercase();
+        let bytes = <sha256::Hash as BitcoinHash>::hash(unit_str.as_bytes());
+        // Take the first 4 bytes and convert to u32 (big endian) make sure the integer
+        u32::from_be_bytes([bytes[0], bytes[1], bytes[2], bytes[3]]) & !(1 << 31)
+    }
 }


I am concerned that this is not a hardened solution. It would derive different derivation indexes for currency unit strings that are almost but not quite the same:

leading and trailing whitespace

capitalization differences

functionally equivalent unicode character combinations

The version in this PR also doesn't include a reserved range for the existing hard coded currency units. Over a long enough timeframe this will introduce derivation path conflicts with sats, millisats, USD, and EUR.

As written in this PR, the hashed_derivation_index function would open the door to bugs and exploits. Resolving those issues in a backward compatible way would be more complex and difficult than just addressing the issues now before any hashed currency unit derivation paths are introduced in the wild.

I just removed the draft setting on #1181. I will rebase it to the latest commit on main. I think we should get 1181 merged and rebase this commit on top of it. Y'all should also review the NUT spec PR: cashubtc/nuts#292. Once I get some feedback on it I would be happy to open a PR to nutshell to get the spec finalized.

lescuer97 · 2025-12-04T12:15:05Z

@vnprc @crodas could you take a look at this pr again pls. I think it is ready now

lescuer97 · 2026-01-01T20:44:01Z

@crodas could you re run the test? I think there was an error

thesimplekid · 2026-01-01T23:08:01Z

@crodas could you re run the test? I think there was an error

Its not an error. Cargo.lock file needs to be updated since you added a dep.

thesimplekid · 2026-01-06T11:55:23Z

@crodas tests are now passing can you review?

crodas · 2026-01-06T11:57:34Z

Absolutely!

lescuer97 · 2026-01-14T13:09:33Z

@thesimplekid rebased.

I will do some final tests around compatibility and it should be ready

lescuer97 · 2026-01-16T12:24:00Z

demo of the spec:

remote_signer_demo_compresed_free_convert.mp4

lescuer97 changed the title ~~bring signatory up to date with the signatory spec~~ bring signatory up to date with the remote signer spec Nov 8, 2025

lescuer97 force-pushed the signatory_up_to_spec branch from e44ec92 to 4e26d87 Compare November 8, 2025 19:21

thesimplekid requested a review from crodas November 11, 2025 15:59

thesimplekid added this to the 0.14.0 milestone Nov 12, 2025

thesimplekid reviewed Nov 12, 2025

View reviewed changes

thesimplekid assigned crodas Nov 13, 2025

thesimplekid reviewed Nov 14, 2025

View reviewed changes

vnprc reviewed Nov 17, 2025

View reviewed changes

thesimplekid removed this from the 0.14.0 milestone Nov 17, 2025

lescuer97 force-pushed the signatory_up_to_spec branch from 891ac38 to 473d6c1 Compare December 3, 2025 14:13

lescuer97 requested a review from vnprc December 5, 2025 12:43

crodas previously approved these changes Jan 9, 2026

View reviewed changes

thesimplekid mentioned this pull request Jan 11, 2026

Deterministic Currency Unit Derivation Paths #1181

Closed

2 tasks

lescuer97 mentioned this pull request Jan 11, 2026

NUT-XX - Mint Remote signer standard cashubtc/nuts#250

Open

3 tasks

lescuer97 dismissed crodas’s stale review via 910582e January 13, 2026 14:53

lescuer97 force-pushed the signatory_up_to_spec branch from ce313d4 to 910582e Compare January 13, 2026 14:53

crodas previously approved these changes Jan 14, 2026

View reviewed changes

lescuer97 added 2 commits January 15, 2026 11:08

rebase

5f48cd5

fix: tests

9769f00

lescuer97 dismissed crodas’s stale review via 9769f00 January 15, 2026 10:09

lescuer97 force-pushed the signatory_up_to_spec branch from 910582e to 9769f00 Compare January 15, 2026 10:09

ye0man added this to CDK Jan 15, 2026

github-project-automation bot moved this to Backlog in CDK Jan 15, 2026

bring signatory up to date with the remote signer spec #1257

Are you sure you want to change the base?

bring signatory up to date with the remote signer spec #1257

Uh oh!

Conversation

lescuer97 commented Nov 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Notes to the reviewers

Suggested CHANGELOG Updates

CHANGED

ADDED

REMOVED

FIXED

Checklist

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lescuer97 Nov 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

thesimplekid commented Nov 17, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lescuer97 commented Dec 4, 2025

Uh oh!

lescuer97 commented Jan 1, 2026

Uh oh!

thesimplekid commented Jan 1, 2026

Uh oh!

thesimplekid commented Jan 6, 2026

Uh oh!

crodas commented Jan 6, 2026

Uh oh!

lescuer97 commented Jan 14, 2026

Uh oh!

lescuer97 commented Jan 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

lescuer97 commented Nov 8, 2025 •

edited

Loading

lescuer97 Nov 14, 2025 •

edited

Loading

lescuer97 commented Jan 16, 2026 •

edited

Loading