Contract Source Code:

// SPDX-License-Identifier: LGPL-3.0-only
pragma solidity >=0.7.0 <0.9.0;

/// @title Multi Send - Allows to batch multiple transactions into one.
/// @author Nick Dodson - <[email protected]>
/// @author Gonçalo Sá - <[email protected]>
/// @author Stefan George - <[email protected]>
/// @author Richard Meissner - <[email protected]>
contract MultiSend {
    address private immutable multisendSingleton;

    constructor() {
        multisendSingleton = address(this);
    }

    /// @dev Sends multiple transactions and reverts all if one fails.
    /// @param transactions Encoded transactions. Each transaction is encoded as a packed bytes of
    ///                     operation as a uint8 with 0 for a call or 1 for a delegatecall (=> 1 byte),
    ///                     to as a address (=> 20 bytes),
    ///                     value as a uint256 (=> 32 bytes),
    ///                     data length as a uint256 (=> 32 bytes),
    ///                     data as bytes.
    ///                     see abi.encodePacked for more information on packed encoding
    /// @notice This method is payable as delegatecalls keep the msg.value from the previous call
    ///         If the calling method (e.g. execTransaction) received ETH this would revert otherwise
    function multiSend(bytes memory transactions) public payable {
        require(address(this) != multisendSingleton, "MultiSend should only be called via delegatecall");
        // solhint-disable-next-line no-inline-assembly
        assembly {
            let length := mload(transactions)
            let i := 0x20
            for {
                // Pre block is not used in "while mode"
            } lt(i, length) {
                // Post block is not used in "while mode"
            } {
                // First byte of the data is the operation.
                // We shift by 248 bits (256 - 8 [operation byte]) it right since mload will always load 32 bytes (a word).
                // This will also zero out unused data.
                let operation := shr(0xf8, mload(add(transactions, i)))
                // We offset the load address by 1 byte (operation byte)
                // We shift it right by 96 bits (256 - 160 [20 address bytes]) to right-align the data and zero out unused data.
                let to := shr(0x60, mload(add(transactions, add(i, 0x01))))
                // We offset the load address by 21 byte (operation byte + 20 address bytes)
                let value := mload(add(transactions, add(i, 0x15)))
                // We offset the load address by 53 byte (operation byte + 20 address bytes + 32 value bytes)
                let dataLength := mload(add(transactions, add(i, 0x35)))
                // We offset the load address by 85 byte (operation byte + 20 address bytes + 32 value bytes + 32 data length bytes)
                let data := add(transactions, add(i, 0x55))
                let success := 0
                switch operation
                    case 0 {
                        success := call(gas(), to, value, data, dataLength, 0, 0)
                    }
                    case 1 {
                        success := delegatecall(gas(), to, data, dataLength, 0, 0)
                    }
                if eq(success, 0) {
                    revert(0, 0)
                }
                // Next entry starts at 85 byte + data length
                i := add(i, add(0x55, dataLength))
            }
        }
    }
}