feat(sdk-coin-sui): add MPCv2 support to recoverConsolidations - #9496

WCI-1234

AI-generated review — auto-generated by Claude Code (wallet-platform:review-pr). Please use your own judgment before acting on any findings.

Phase 1 — Summary

This PR adds MPCv2 support to Sui.recoverConsolidations() by detecting EDDSA signing material once before the scan loop and threading it as an optional precomputedMaterial parameter through recover() → recoverSuiToken() → signRecoveryTransaction(). Previously each loop iteration would independently decrypt the keycard via getEddsaSigningMaterial. The change is purely additive — existing v1 and unsigned sweep paths are unaffected.

Files changed: 2 | +191 / −8

• modules/sdk-coin-sui/src/sui.ts — production change
• modules/sdk-coin-sui/test/unit/sui.ts — 4 new tests in new Build Consolidation Recoveries (MPCv2): describe block

flowchart TD
    A[recoverConsolidations] -->|"userKey && walletPassphrase?"| B{detect once}
    B -->|yes| C[getEddsaSigningMaterial → signingMaterial]
    B -->|no| D[signingMaterial = undefined]
    C --> E[scan loop: idx = startingScanIndex..endingScanIndex]
    D --> E
    E -->|per iteration| F["recover(recoverParams, signingMaterial)"]
    F -->|token branch| G["recoverSuiToken(..., precomputedMaterial)"]
    F -->|SUI branch| H["signRecoveryTransaction(..., precomputedMaterial)"]
    G --> H
    H --> I["precomputedMaterial ?? getEddsaSigningMaterial(...)"]
    I --> J[sign & return tx]

Phase 2 — Ticket Alignment

Ticket: WCI-1234

PR title: feat(sdk-coin-sui): add MPCv2 support to recoverConsolidations

The implementation directly addresses what the ticket implies: extend MPCv2 signed hot recovery (landed in part 1, #9493) to the consolidation sweep path, matching the DOT pattern (WCI-1236). Alignment is strong.

Phase 3 — Index Analysis

No new database queries introduced.

Phase 4 — Type Safety

Double casts found (as unknown as) — all in test code for spying on private methods:

These patterns are acceptable for sinon spying on private methods and are consistent with existing test patterns in the file. No as any, @ts-ignore, or production-code double casts found.

One minor note: typing getEddsaSigningMaterial as unknown in the spy cast means the spy object loses its inferred call signature — calledOnce still works because it comes from sinon.SinonSpy, but it's slightly weaker than casting to the actual method type. Not a blocker.

Phase 5 — MPCv2 Correctness & Test Quality

Correctness

Test Quality

Consistency vs DOT (#9486) and ADA (#9497)

Summary of Findings

No blocking issues. One minor test-quality flag:

• ⚠️ v1 regression spy assertion (sandBox.assert.called(getTSSSignatureSpy)) is weaker than it could be — consider calledTwice if the 2-address scan (indexes 1–2) is deterministic, to catch any future regression that skips an iteration.

Everything else is clean: the detection-once pattern is correctly implemented, all four test scenarios are covered, and the unsigned sweep path is properly gated.

Generated by [wallet-platform:review-pr] via Claude Code