Raven: Mining Ethereum Defensive Patterns

Published:Dec 27, 2025 14:47
1 min read
ArXiv

Analysis

This paper introduces Raven, a framework for identifying and categorizing defensive patterns in Ethereum smart contracts by analyzing reverted transactions. It's significant because it leverages the 'failures' (reverted transactions) as a positive signal of active defenses, offering a novel approach to security research. The use of a BERT-based model for embedding and clustering invariants is a key technical contribution, and the discovery of new invariant categories demonstrates the practical value of the approach.

Reference

Raven uncovers six new invariant categories absent from existing invariant catalogs, including feature toggles, replay prevention, proof/signature verification, counters, caller-provided slippage thresholds, and allow/ban/bot lists.