secp256r1-verify double-hash behavior on mainnet vs simnet

[Rapha-btc]

No description provided.

[Rapha-btc]

maybe not necessary after all

[Rapha-btc]

Question: secp256r1-verify double-hash behavior on mainnet vs simnet

Description:

Tests pass with single-hash in secp256r1-utils.ts:

const createFinalMessageHash = (prefix, domainHash, messageHash) => {
  return createHash("sha256")
    .update(Buffer.concat([prefix, domainHash, messageHash]))
    .digest();  // single hash
};

However, per [stacks-core PR #6763](stacks-network/stacks-core#6763), mainnet's secp256r1-verify has a double-hash bug (fixed in upcoming Clarity 5).

Question for @obycode @friedger:

Does simnet in current Clarinet already have the fix, or does mainnet require double-hash in the frontend?

// If mainnet needs double-hash:
const createFinalMessageHash = (prefix, domainHash, messageHash) => {
  const singleHash = createHash("sha256")
    .update(Buffer.concat([prefix, domainHash, messageHash]))
    .digest();
  return createHash("sha256").update(singleHash).digest();
};

Need to confirm before deploying passkey auth to production.

[Rapha-btc]

Yes, exactly.

When Clarity 5 ships with Brice's fix:

What	Change needed?
CSW Clarity contracts	❌ No changes
Frontend/TypeScript	✅ Update sign() to not pre-hash

The Clarity auth helpers just build a hash. The secp256r1-verify function is what was buggy (hashing twice). Once fixed, the frontend needs to match by only hashing once.

CSW contracts are future-proof. Only the signing code (frontend/tests) needs updating later.

[Rapha-btc]

Yes — Brice's fix changes secp256r1-verify to not hash, so the frontend will stop pre-hashing and that's it.