fix(psbt): route addOutput unknownKeyVals to output map - #21

What

• Fix `Psbt.addOutput()` in `ts_src/lib/psbt.ts:159`: change
  `this.addUnknownKeyValToInput(outputIndex, keyVal)` →
  `this.addUnknownKeyValToOutput(outputIndex, keyVal)`.
• Add two regression tests in `ts_src/tests/addInputOutput.ts`:
  • Verifies unknown key-values land on `outputs[N].unknownKeyVals`
    (not `inputs[N].unknownKeyVals`) and survive a `toHex()`/`fromHex()` round-trip.
  • Verifies the crash `"No input #N"` no longer occurs when the number
    of outputs exceeds the number of inputs (e.g. 1-input / 2-output transactions).

Why

`addOutput()` was calling the input variant of `addUnknownKeyVal` instead of
the output variant, causing two confirmed failure modes (WCN-1934 / FND-001):

1. Silent corruption when `outputIndex < inputs.length`: the keyval was
   attached to `inputs[outputIndex].unknownKeyVals` and survived
   serialize/parse round-trips in the wrong map. The PSBT proprietary-key
   support added for musig2/MPC coordination data (commit 8a33996) was
   completely broken: co-signers never found the required output metadata,
   causing signing ceremonies to fail with funds stuck.

2. Crash (`"No input #N"`) when `outputIndex >= inputs.length` — the
   common 1-input / 2-output case — making PSBT construction impossible
   whenever `unknownKeyVals` accompany outputs beyond the first.

Additionally, the old code ran the duplicate/type guard against `InputTypes`
(16 entries) instead of `OutputTypes` (6 entries), so output unknown keys with
type byte 6–15 were additionally misrejected. This is addressed in pt2.

Test plan

• `npm run unit` — 387 tests pass (2 new regression tests added)
• `npm run format:ci` — no formatting issues

Stack

This PR is part 1 of 2 in a stack. Review and merge in order:

1. fix(psbt): route addOutput unknownKeyVals to output map #21 — primary routing fix + regression tests (base: `master`) ← you are here
2. fix(utils): use max-enum-value+1 as checkHasKey boundary; reject empty keys #22 — `checkHasKey`/`getEnumLength` correctness fixes (base: `pt1`)

Ticket: WCN-1934