Submitted by BI_security, also found by plasmablocks, CipherSleuths (1, 2, 3, 4), J4X, pkqs90, NentoR, haxatron, rokinot, SBSecurity (1, 2), EV_om (1, 2), trachev, hash (1, 2, 3), SpicyMeatball, Ward, juancito, jasonxiale, hals (1, 2), 0xPsuedoPandit, KingNFT, KupiaSec, Giorgio, deepplus, zaevlad, Tendency, ABAIKUNANBAEV, Hajime, Beepidibop, 0xpiken (1, 2), ravikiranweb3, zzebra83, boringslav, ZdravkoHr, and rvierdiiev (1, 2, 3, 4)
https://github.com/re-nft/smart-contracts/blob/3ddd32455a849c3c6dc3c3aad7a33a6c9b44c291/src/packages/Signer.sol#L151-L151 
https://github.com/re-nft/smart-contracts/blob/3ddd32455a849c3c6dc3c3aad7a33a6c9b44c291/src/packages/Signer.sol#L373-L375 
https://github.com/re-nft/smart-contracts/blob/3ddd32455a849c3c6dc3c3aad7a33a6c9b44c291/src/packages/Signer.sol#L384-L386 
https://github.com/re-nft/smart-contracts/blob/3ddd32455a849c3c6dc3c3aad7a33a6c9b44c291/src/packages/Signer.sol#L232-L238 
https://github.com/re-nft/smart-contracts/blob/3ddd32455a849c3c6dc3c3aad7a33a6c9b44c291/src/policies/Create.sol#L636
Being not EIP712 compliant can lead to issues with integrators and possibly DOS.
The implementation of the hook hash (here) is done incorrectly. hook.extraData is of type bytes which according to EIP712 it is referred to as a dynamic type. Dynamic types must be first hashed with keccak256 to become one 32-byte word before being encoded and hashed together with the typeHash and the other values.
Some TypeHashes are computed by using abi.encode instead of abi.encodePacked (here) which makes the typeHash value and the whole final hash different from the hash that correctly implementing EIP712 entities would get.
The problem with this is that abi.encode-ing strings results in bytes with arbitrary length. In such cases (like the one here) there is a high chance that the bytes will not represent an exact N words in length (X * 32 bytes length) and the data is padded to conform uniformly to 32-byte words. This padding results in an incorrect hash of the typeHash and it will make the digest hash invalid when compared with properly implemented hashes from widely used libraries such as ethers.
Place the following code in any of the tests and run forge test -mt test_EIP712_encoding
This test shows that the rentalOrderTypeHashCorrect is the correct typeHash.
_deriveOrderMetadataHash  constructs the hash incorrectly because _ORDERMETADATATYPEHASH includes uint8 orderType and bytes emittedExtraData (it can be seen here) but these values are not provided below the typeHash (it can be seen here).
This hash is important because it is compared to the zoneHash inside Create.sol:validateOrder  _rentFromZone  _isValidOrderMetadata (link): So if the provided zoneHash from Seaport was generated correctly and it is not the same as the one generated by reNFT, the protocol will not be able to create any rentalOrders resulting in DOS.
In any case, this implementation is not according to EIP712 and either the fields must be included or the _ORDER_METADATA_TYPEHASH must remove uint8 orderType and bytes emittedExtraData
Foundry
Apply the described fixes for each example.
Alec1017 (reNFT) confirmed
reNFT mitigated:
Status: Mitigation confirmed. Full details in reports from juancito, EV_om and sin1st3r__.
