feat(sdk-coin-kaspa): implement Kaspa SDK module

[mohd-kashif]

• Kaspa coin class (UTXO, Schnorr/secp256k1, 8 decimals / sompi)
• TransactionBuilder + TransactionBuilderFactory
• KeyPair (Secp256k1ExtendedKeyPair, kaspa bech32 address derivation)
• Transaction with Blake2b-256 sighash and Schnorr signatures
• Utils: kaspa/kaspatest bech32 address validation
• Full test suite: keyPair, transaction, builder, utils, coin, transactionFlow tests
• .mocharc.yml and eslint config

Jira: CECHO-388

[parasgarg-bitgo]

If Kaspa follows UTXO model, why are we not extending abstract-UTXO module for it?

[mohd-kashif]

If Kaspa follows UTXO model, why are we not extending abstract-UTXO module for it?

Kaspa shares only the UTXO accounting model — everything else is different:

Blake2b-256 hashing, not SHA-256
Schnorr-only with x-only pubkeys and a Kaspa-specific OP_CHECKSIG_SCHNORR (0xAB) opcode — not Bitcoin's Taproot Schnorr
cashaddr-style bech32 with a custom polynomial and kaspa:/kaspatest: prefix — incompatible with Bitcoin bech32
Kaspa-specific tx fields (subnetworkId, gas, payload, per-input sigOpCount) and its own BIP-143-like sighash with different pre-image layout
BlockDAG (GHOSTDAG) consensus, no PSBT standard, no Bitcoin-compatible wire format
Extending abstract-utxo would mean overriding nearly every method and shoe-horning Kaspa into utxo-lib, which doesn't know how to parse/sign Kaspa transactions. The clean separation — extend BaseCoin / BaseTransactionBuilder / Secp256k1ExtendedKeyPair — is what the codebase already does for other non-Bitcoin UTXO-ish chains (e.g., Cardano in sdk-coin-ada).