Transfer-Hook Interface
Example program
Here is an example program that only implements the required "execute" instruction, assuming that the proper account data is already written to the appropriate program-derived address defined by the interface.
use ;
Motivation
Token creators may need more control over how their token is transferred. The most prominent use case revolves around NFT royalties. Whenever a token is moved, the creator should be entitled to royalties, but due to the design of the current token program, it's impossible to stop a transfer at the protocol level.
Current solutions typically resort to perpetually freezing tokens, which requires a whole proxy layer to interact with the token. Wallets and marketplaces need to be aware of the proxy layer in order to properly use the token.
Worse still, different royalty systems have different proxy layers for using their token. All in all, these systems harm composability and make development harder.
Solution
To improve the situation, Token-2022 introduces the concept of the transfer-hook interface and extension. A token creator must develop and deploy a program that implements the interface and then configure their token mint to use their program.
During transfer, Token-2022 calls into the program with the accounts specified at a well-defined program-derived address for that mint and program id. This call happens after all other transfer logic, so the accounts reflect the end state of the transfer.
How to Use
Developers must implement the Execute
instruction, and optionally the
InitializeExtraAccountMetaList
instruction to write the required additional account
pubkeys into the program-derived address defined by the mint and program id.
Note: it's technically not required to implement InitializeExtraAccountMetaList
at that instruction descriminator. Your program may implement multiple interfaces,
so any other instruction in your program can create the account at the program-derived
address!
When your program stores configurations for extra required accounts in the
well-defined program-derived address, it's possible to send an instruction -
such as Execute
(transfer) - to your program with only accounts required
for the interface instruction, and all extra required accounts are
automatically resolved!
Account Resolution
Implementers of the transfer-hook interface are encouraged to make use of the spl-tlv-account-resolution library to manage the additional required accounts for their transfer hook program.
TLV Account Resolution is capable of powering on-chain account resolution when an instruction that requires extra accounts is invoked. Read more about how account resolution works in the repository's README file.
An Example
You have created a DAO to govern a community. Your DAO's authority is a multisig account, and you want to ensure that any transfer of your token is approved by the DAO. You also want to make sure that someone who intends to transfer your token has the proper permissions to do so.
Let's assume the DAO multisig has some fixed address. And let's assume that
in order to have the can_transfer
permission, a user must have this
dynamic program-derived address associated with their wallet via the
following seeds: "can_transfer" + <wallet_address>
.
Using the transfer-hook interface, you can store these configurations in the well-defined program-derived address for your mint and program id.
When a user attempts to transfer your token, they might provide to Token-2022:
Token-2022 will then call into your program, resolving the extra required accounts automatically from your stored configurations, to result in the following accounts being provided to your program:
Utilities
The spl-transfer-hook-interface
library provides offchain and onchain helpers
for resolving the additional accounts required. See
invoke.rs
for usage on-chain, and
offchain.rs
for fetching the additional required account metas with any async off-chain client
like BanksClient
or RpcClient
.