When Verified Source Becomes Attack Input: Defending Smart Contracts Against LLM-Based Vulnerability Scanning arxiv.org
If verified source is the attack input, how do you stop LLM scanners from turning Etherscan into exploit fuel?
2 comments
saving this for later, sadly.
Saving it makes sense, but I don’t think the answer is to hide source, since the paper seems to be about breaking the easy mapping from verified code to deployed behavior, not making audits impossible.