This repository introduces an example showing how Violet's Identity and Compliance system can be used to add a permission layer on top of a Uniswap V4 pool, using Hooks. It relies on the VioletID registry, which contains a mapping from an address to the attributes/statuses associated with the entity registered with Violet controlling that address. To learn more about it, please see this page from Violet's docs.
The VioletID registry is live on Ethereum Mainnet at 0xfFcDa323597A90af8F7A0eCBE4ef14e9b30e4a9e.
As shown in VioletHooksExample
, specific checks can be set for each pool such that:
- Only addresses with specific status(es) can interact with a pool
- And/or only addresses without specific status(es) can interact with a pool
When a hook is called, the Hooks contract in turn calls the VioletID registry and compare the VioletID statuses of the user against the requirements set for the pool involved in the transaction.
Although the first field received as parameter in a V4 hook is an address called sender
, this address might not correspond to the user's address, the one that ultimately should be checked against the VioletID Registry. The reason being that the PoolManager will call the hook contract passing msg.sender
(see here for a swap for example).
However, just like in V3, it is likely that the address calling the PoolManager
is not the end user but likely another contract such as a SwapRouter
which sits in between.
In order to get the end users address then, the tests uses an approach passing it as hookData
.
hookData
is arbitrary calldata that can be passed along when calling the PoolManager
, therefore in order to guarantee a correct identity check, it shifts the responsibility to the router contract to ensure that the end user's address initiating the interaction with a pool is correctly set as part of hookData
. This is done in test/utils/routers
.
If a router contract doesn't correctly set the end user's address in hookData, but for example a hardcoded address of a fully compliant entity instead, this would circumvent the checks defined in the hooks - they will always pass regardless of the address which initiated the transaction. Therefore, the Hooks contract example here (VioletHooksExample
) also contains logic to whitelist router contracts (received as sender in hooks).
requires foundry
forge install
forge test
Because v4 exceeds the bytecode limit of Ethereum and it's business licensed, we can only deploy & test hooks on anvil.
# start anvil, with a larger code limit
anvil --code-size-limit 30000
# in a new terminal
forge script script/Counter.s.sol \
--rpc-url http://localhost:8545 \
--private-key 0xac0974bec39a17e36ba4a6b4d238ff944bacb478cbed5efcae784d7bf4f2ff80 \
--code-size-limit 30000 \
--broadcast
When installing dependencies with forge install
, Github may throw a Permission Denied
error
Typically caused by missing Github SSH keys, and can be resolved by following the steps here
Or adding the keys to your ssh-agent, if you have already uploaded SSH keys
Hook deployment failures are caused by incorrect flags or incorrect salt mining
- Verify the flags are in agreement:
getHookCalls()
returns the correct flagsflags
provided toHookMiner.find(...)
- In obscure cases where you're deploying multiple hooks (with the same flags), try setting
seed=1000
forHookMiner.find
- Verify salt mining is correct:
- In forge test: the deployer for:
new Hook{salt: salt}(...)
andHookMiner.find(deployer, ...)
are the same. This will beaddress(this)
. If usingvm.prank
, the deployer will be the pranking address - In forge script: the deployer must be the CREATE2 Proxy:
0x4e59b44847b379578588920cA78FbF26c0B4956C
- If anvil does not have the CREATE2 deployer, your foundry may be out of date. You can update it with
foundryup
- If anvil does not have the CREATE2 deployer, your foundry may be out of date. You can update it with
- In forge test: the deployer for:
Additional resources:
v4-periphery contains advanced hook implementations that serve as a great reference