Research, Reports, And Releases

AI coding assistants can scaffold Solidity, draft tests, and explain concepts in seconds, but they also reproduce the insecure patterns in their training data. We look at where AI helps smart contract work, where it falls short, and how to keep human review and verification in control of what ships.

Ethereum has come through four major operational cases, including the Merge, without permanent finality loss, and with no single operator or SLA behind it. This report sets out eight dimensions institutions can use to assess a public network's operational resilience, applies them to Ethereum's record, and names what still need to be monitored.

Nethermind Security audited Holonym's Aztec Bridge, the cross-chain system that combines Proof of Humanity credentials with Aztec's privacy features. The review uncovered signature replay and state update issues across Ethereum and Aztec contracts, all resolved before mainnet deployment.

Nethermind ran an AI security pilot on already-audited Lido Core, roughly 19,000 of 31,000 LoC on Ethereum. Researchers and AI engineers built protocol-specific context that customized AuditAgent and AgentArena, and pairing both engines with researchers widened coverage and cut false-positive rates.