[OttoAllmendinger]

Add a PoX-5 lockup validator to the descriptor policy (offlineVault and
prod getPolicyForEnv), accepting only canonical lockup descriptors whose
staker keys are the wallet keys. Template validators now return false for
descriptors they cannot parse instead of throwing. Half-signing applies
the PoX-5 staking spend policy (nLockTime, non-final sequence, exactly
one lockup input). The descriptor explanation now includes locktime.

[linear-code]

WAL-2156

[github-actions]

⚠️ Unit tests are failing on Node 26.x (Current release line, non-blocking). This is not an LTS version yet, so it does not block merge, but it signals an incompatibility to fix before Node 26.x becomes LTS.

View run

[Doddanna17]

LGTM — clean, security-conscious change with exactly the right tests.

What I checked:

• Lockup validator is correctly strict. isCanonicalPox5Lockup rebuilds the descriptor from the wallet keys and accepts only when canonical.toString() === descriptor.toString(), which binds the staker keys to the wallet keys and rejects non-canonical encodings. getValidatorPox5Wallet = eachOneOf(Wsh2Of3, pox5Lockup) cleanly lets a wallet carry both descriptor kinds.
• Spend policy is right. assertPox5SpendPolicy no-ops when there are no PoX-5 inputs; otherwise requires exactly one lockup input, branch locktime (early-exit rejected), nLockTime ≥ max(unlockHeights), and a non-final sequence; getMatchedTransactionInput also verifies the PSBT input's witnessUtxo script equals the descriptor's scriptPubkey.
• Template validators return false instead of throwing for unparseable descriptors — needed for the eachOneOf fallthrough and doesn't weaken acceptance (still rejects). Covered by a test.
• Tests hit the real attack vectors: foreign/permuted staker keys, malformed + non-canonical encodings, nLockTime below unlock / as a timestamp, final sequence, >1 lockup input, plus getPolicyForEnv(prod) accept/reject and a spend-policy compat test for the code moved from utxo-staking.

Non-blocking:

• names.ts maps tbtcstx/tbtcstxprivate1 → tbtcreg and tbtcstxsignet → tbtcsig; btc/tbtcbgsig pass through (valid wasm names) — no gap.
• API surface changed (new policy exports + deprecated utxo-staking re-exports) — downstream must pick up the matching SDK version.
• tryParseDescriptor now hides "not a template" vs "genuinely malformed" — fine here, just noting.

CI green on 1afbaaec. Approving.