#41899 [BC-Insight] NatSpec of several functions in `ethereum.move` is wrong

Submitted on Mar 19th 2025 at 09:07:07 UTC by @avoloder for Attackathon | Movement Labs

• Report ID: #41899

• Report Type: Blockchain/DLT

• Report severity: Insight

• Target: https://github.com/immunefi-team/attackathon-movement-aptos-core/tree/main

• Impacts:

Description

Brief/Intro

NatSpec comments in several functions are wrong with regard when the functions should abort (aptos-move/framework/aptos-framework /sources/ethereum.move)

Vulnerability Details

Functions ethereum_address_no_eip55() and ethereum_address_20_bytes() contain comments that they are expected to abort if the address does not conform to EIP-55 standards, which is not true.

ethereum_address_no_eip55() calls assert_40_char_hex() which checks if the address is a nonzero 40-character hexadecimal string, but not if the address is compliant with EIP-55 standard.

ethereum_address_20_bytes() only checks if the vector's length is 20, but also does not check if the address is a valid EIP-55 address.

Impact Details

There is no security impact here, only documentation issue, which might cause confusion as to how the function should behave

References

https://github.com/immunefi-team/attackathon-movement-aptos-core/blob/627b4f9e0b63c33746fa5dae6cd672cbee3d8631/aptos-move/framework/aptos-framework/sources/ethereum.move#L31-L39

https://github.com/immunefi-team/attackathon-movement-aptos-core/blob/627b4f9e0b63c33746fa5dae6cd672cbee3d8631/aptos-move/framework/aptos-framework/sources/ethereum.move#L41-L49

Proof of Concept

Aligning NatSpec with the function's behavior prevents confusion about what the function is supposed to do, i.e., it clearly describes the function's expectations.

It is also a valuable enhancement for the following reasons:

• In collaborative development, it might be unclear whether these functions should abort if they are/are not EIP-55 compliant.

• It helps eliminate false assumptions that the function will abort if it’s not EIP-55 compliant, reducing the risk of errors elsewhere in the code