Adds Babbage/Conway-era reference-script attachment to wallet_send
and wallet_send_unsigned. The output can now carry any combination
of {native assets, inline datum, reference script}.
Why: deploying Plutus validators / minting policies as on-chain
reference UTxOs is the standard Cardano dApp pattern (Agora, Liqwid,
SundaeSwap all do this). Without it every spend or mint that uses
a script has to inline-witness the full CBOR — kilobytes per tx and
quadratic with tx size. Reference scripts let downstream txs witness
via `read_only_input` for ~32 bytes overhead.
API surface:
aldabra-core:
- new `ReferenceScriptSpec<'a> { kind: ScriptKind, cbor: &'a [u8] }`
- `ScriptKind` re-exported from pallas_txbuilder (PlutusV1/V2/V3/Native)
- new `build_signed_payment_extras(...)` and
`build_unsigned_payment_extras(...)` — supersets of the existing
`_with_assets` functions; take both `to_inline_datum_cbor` and
`to_reference_script` Options
- existing `_with_assets` functions kept as thin wrappers that pass
None for ref script — backwards compatible
- internal `output_with_assets`, `prepare_payment`, and
`build_staging_with_fee` thread the new ref-script Option through
aldabra-mcp:
- `SendArgs` and `UnsignedSendArgs` gain
`reference_script_cbor_hex: Option<String>` and
`reference_script_kind: Option<String>`
- Both must be set or both omitted; mismatched returns a clean error
- `parse_script_kind` helper — case-insensitive, accepts
PlutusV1/V2/V3/Native (plus shortcut V1/V2/V3)
Reference scripts are intentionally never attached to change outputs.
The change goes back to the wallet's own address, where a script
attachment would lock value into a publicized script that we'd then
have to spend BACK out — pointless. Ref-script attachment is only
on the recipient (`to`) output.
This unblocks Track B-fast step 3 of the preprod DAO bringup —
deploying the 11 Agora script bytecodes (governor / stakes /
proposal / treasury / mutate / noOp / treasuryWithdrawal validators
+ GST / StakeST / ProposalST / GAT minting policies) as reference
UTxOs against Kayos's preprod wallet.
No new tests in this commit — Phase 2 (Plutus-policy mint with
custom output) lands in a follow-up that will exercise this path
end-to-end against a real chain submit on preprod.
|
||
|---|---|---|
| .cargo | ||
| crates | ||
| docs | ||
| .dockerignore | ||
| .gitignore | ||
| Cargo.lock | ||
| Cargo.toml | ||
| Dockerfile | ||
| LICENSE | ||
| README.md | ||
| ROADMAP.md | ||
aldabra
Rust-native Cardano lite wallet with an MCP-server interface — built for LLM-first usage (send, receive, mint, Plutus interaction).
Status: Phase 1 scaffold (2026-05-04). Compiles, structure in place, real wallet primitives still landing. See
ROADMAP.md.
Why
The existing Cardano MCP servers are either read-only doc gateways (Jimmyh-world/Cardano_MCP) or built on Blockfrost (web3-mcp) which is a centralized API we deliberately don't depend on. Sulkta runs its own Koios + Ogmios endpoints on Rackham; we want a wallet that talks directly to those.
Also: it's the first Sulkta Rust project — useful as a workout for
crafting-table's Rust toolchain (per Sulkta-Coop/lucy-infra
spec-crafting-table.md).
Architecture
Three crates in a Cargo workspace:
| Crate | Responsibility |
|---|---|
aldabra-core |
Pure crypto + types. Mnemonic → root key (CIP-3), root → payment + stake key (CIP-1852), address construction, signing. No I/O, no network. This is the security boundary. |
aldabra-chain |
Pluggable backends for chain queries. ChainBackend trait, with Koios as the phase-1 implementation. Ogmios + submission paths in phase 2. |
aldabra-mcp |
Binary. MCP server speaking stdio. Glues core + chain together, exposes tools to the LLM client. |
┌─────────────────────────────┐
LLM client │ aldabra-mcp (bin) │ stdio
─────────► │ tool handlers, lifecycle │ ────►
└──────────┬──────────────────┘
│
┌────────┴────────┐
▼ ▼
┌──────────────┐ ┌──────────────┐
│ aldabra-core │ │ aldabra-chain │
│ keys, sign │ │ Koios/Ogmios │
└──────────────┘ └──────────────┘
MCP tools (target)
Phase 1:
wallet.address— derived base address at account 0, index 0wallet.balance— ADA + native asset balance at the wallet's addresswallet.utxos— list UTXOs
Phase 2:
wallet.send— build, sign, submit a payment (ADA or native)wallet.tx_status— poll a submitted tx hash
Phase 3:
wallet.mint— mint a CIP-25 / CIP-68 native assetwallet.policy.create— generate a policy script (timelock, multisig)
Phase 4:
wallet.script.attach— attach an inline datum + reference scriptwallet.script.spend— spend a Plutus-locked UTXO with redeemerwallet.stake.delegate— delegate to a pool
Build
# Local (requires rustc 1.75+)
cargo build --release
# Through crafting-table (preferred — validates the toolchain there)
crafting-table build aldabra
Run
# Direct invocation (smoke test only — does nothing useful in phase 1)
./target/release/aldabra
# As an MCP server registered with Claude Code:
# add to ~/.claude.json:
# "aldabra": {
# "command": "/path/to/aldabra",
# "env": {
# "ALDABRA_DATA": "/mnt/cache/appdata/aldabra"
# }
# }
Security model
- Mnemonic source: interactive bootstrap on first run, paste once, encrypted at rest with age. Never written to disk in plaintext.
- Derived keys: in-memory only,
ZeroizeOnDropon every container. - Network exposure: stdio MCP transport — never opens a TCP socket. Only the spawning client process can reach it.
- Multi-network: mainnet by default, but
--network preview/--network preprodfor testing without real ADA.
See also
ROADMAP.md— phased buildoutdocs/architecture.md— deeper design notes- txpipe/pallas — the Rust Cardano building blocks we depend on
- Emurgo/cardano-serialization-lib — reference TX builder if pallas-txbuilder doesn't cover something
- modelcontextprotocol/rust-sdk — rmcp, the Rust MCP server SDK we use