Testing current master with lethekit

## Abstract

Testing 7604c8089008896bf86c7c18c800aeb86fe5f83e

### Test vector 1
```bash
seed_lethekit=$(seedtool --in ur ur:crypto-seed/oyadgdvlpfssfwmkztcebbnyzowkspnljlrhdkztykahhf)

keytool \
--seed $seed_lethekit \
--address-index '*' \
output-descriptor
```

result OK, matches with lethekit:
```bash
wpkh([b238a845/84h/0h/0h]xpub6D1GQL3S3qryEx5gQccdVB5Fy1ePkeDGU8bMmW1GKfDQvtBsbJ6FaaudsBphw2WwtjYyJ4ARR6Mh9ZRTTKqro6eZww2m6SVxG7WuY2DALVp/0/*)
```

### Test vector 2

```bash
# address wpkh not supported only pkh and sh?
for (( i=0; i <= 9; ++i ))
do
keytool \
--seed 1eb5338edac6fae54cbb172091ae6c1a \
--address-index $i \
address-index \
full-address-derivation-path \
address-pkh
done
```

Result:
```bash
0
c1859926/84h/0h/0h/0/0
17u4HStkGWxJ5PtNqzCMEKRzaPJ8h35eBc
1
c1859926/84h/0h/0h/0/1
1JyxrbHYS7ZGhq68JewhJAMWka38Bv8tFw
2
```

Here I noticed no support for segwit addresses? Also, the output shows segwit derivation path (84h) when `address-pkh` specified (`44h`) -> Not OK.

---

You can generate more test vectors for a particular seed (e.g. lethekit_seed/mnemonic: `together mail awful cradle scrub apart hip leader silk slice unusual embark`) in https://iancoleman.io/bip39/

Start by reproducing test vector 2 with keytool's address-index, full-address-derivation-path, and address-pkh entry points, then trace how those commands select derivation paths and address formats. Compare the output with the reported 44h expectation and segwit observation; done means the behavior is clarified or corrected and covered by repeatable test vectors.

索引モデルが issue の本文から書いたものです。