#36577 psbt: accept taproot ALL sigs with sighash unset

Relaxes signature verification in `SignPSBTInput` to accept 65-byte Taproot signatures ending in `SIGHASH_ALL` when neither `PSBT_IN_SIGHASH_TYPE` nor an explicit `sighash_type` option is provided. Previously, unset sighash types strictly assumed 64-byte `SIGHASH_DEFAULT` signatures and rejected any existing 65-byte signatures with `SIGHASH_MISMATCH`. Mismatches remain rejected if an explicit sighash type or field is set, or if the 65-byte signature uses sighashes other than `SIGHASH_ALL`.

Problem: Taproot signers sometimes produce 65-byte signatures explicitly specifying `SIGHASH_ALL` while leaving `PSBT_IN_SIGHASH_TYPE` unset. When a subsequent signer or finalizer processes the PSBT without an explicit sighash parameter, `SignPSBTInput` falls back to `SIGHASH_DEFAULT` and rejects valid 65-byte `SIGHASH_ALL` signatures as invalid, preventing the transaction from being signed further or finalized.

Evidence of importance:

• Fixes issue #36507: 'PSBT: SignPSBTInput rejects valid 65-byte taproot script signatures when PSBT_IN_SIGHASH_TYPE is absent'

Risk of a mistake: low (wallet, rpc_api). A flaw in PSBT input signature checks could mistakenly allow an unintended sighash type or break PSBT signing workflows.

Changed since the previous assessment: Unchanged in approach; provides the implementation and accompanying unit tests for the previously described change.

Ready

Nothing known that would invalidate a review now.

Waiting on (from the thread read):

• nothing: nothing pending

Strong

substantive support, no open objection (Husteem)

Thread notes: Added Husteem's Approach ACK / Tested ACK. No objections or suggestions were raised.

Objections, suggestions, and questions: none recorded.

Support:

• Yudis-bit (none) (Tested ACK, 2026-10-20) [tested]: Tested ACK 1291d0e389 on Windows. [not substantive]
• Husteem (none) (Approach ACK, 2026-10-22) [both: psbt_tests, rpc_psbt.py]: accepting 65-byte Taproot signatures ending in SIGHASH_ALL when sighash_type is unset is safe because it commits to the identical BIP-341 digest as SIGHASH_DEFAULT, while preserving safety checks for other variants

Participants:

• Yudis-bit (none, support, 1 statements 2026-10-20 to 2026-10-20): tested ACK on Windows
• Husteem (none, support, 1 statements 2026-10-22 to 2026-10-22): Approach ACK / Tested ACK with rationale and testing on Ubuntu

79 lines under test/bench/ci.

• src/test/psbt_tests.cpp +79/-0
• src/psbt.cpp +8/-2

The short text the ranking pass compares across PRs.

Relaxes SignPSBTInput to accept 65-byte Taproot signatures ending in SIGHASH_ALL when no sighash type is explicitly specified by the caller or PSBT input. Previously, omitting the sighash field caused SignPSBTInput to assume SIGHASH_DEFAULT and strictly demand 64-byte signatures, causing valid PSBTs with 65-byte SIGHASH_ALL signatures to fail validation and blocking them from being signed further or finalized. This fixes issue #36507.

P3 · bug fix (ranking pass: Moved up to sit with the other PSBT signing interoperability fixes (#36039, #35747).)

Fixes an interoperability bug where SignPSBTInput rejects valid 65-byte Taproot SIGHASH_ALL signatures when PSBT_IN_SIGHASH_TYPE is unset. As noted in issue #36507, this incorrectly triggers SIGHASH_MISMATCH and blocks downstream signing or finalization.

Membership: src/psbt.cpp implements core PSBT handling and SignPSBTInput used by the wallet to process and sign transactions.

Factors: security/stability 0, bug 1, performance 0, user value 1, leverage 0