VVE-2020-0001

Earlier today, we received a responsible disclosure of a potential issue from @montyly (security researcher at @trailofbits) for Vyper users who make assumptions about what values certain interface types can return.

Impact

We determined the issue to be mild and unlikely to be exploited, with an easy workaround while the correct resolution is in process. The issue stems from a number of things, which we will detail here.

(1) The ABI Specification is under-defined such that function return type is not always reflected in how you use it

This means that a function which returns uint8 under the hood actually returns a 32 byte integer, making it identical to a function that returns uint256. This allows users to read an interface that returns a uint8 value to be stored into a uint256 variable without any explicit casting or input validation.

(2) Vyper doesn't have uint8 types

When Vyper was originally created, it only had one numeric type, but we added just enough types to be able to work with the majority of ERC interfaces that exist.

Unfortunately, we never added uint8, because it's only majority usage was for ERC20.decimals() as the return type, which isn't reflected in the method ID. Because of (1), it didn't matter that we didn't have these types implemented because you could capture the return value as uint256 and use it just fine.

(3) ERC20.decimals() returns uint8

ERC20.decimal() (which is an optional function) returns a uint8 type. While it was never intentioned to be used directly within a smart contract (hence being optional), someone could easily make the decision to rely on it to perform important functionality within their Vyper smart contract. This might lead to a scenario where an unexpectedly large value (> 255) returned by calling this function (which a malicious contract writer might write) would allow an attacker to manipulate or bypass certain logic depending on this value.

In summary, because of (1), it isn't necessary to have to cast the return value of a function that returns uint8 to uint256, and because of (2) it isn't possible to have the type system protect against this type of error. This could lead to scenarios like (3) where this behavior can be exploited.

Patches

We are currently refactoring our typing system so we can implement all ABI-compliant integer types, but no currently patched version is available that gives users access to the uint8 type.

Workarounds