Logo image
TOPOS: Topological Profiling of On chain Subgraphs for Cross Chain Forensics
Abstract

TOPOS: Topological Profiling of On chain Subgraphs for Cross Chain Forensics

Aravinda S. Rao and Babu Pillai
2026 IEEE International Conference on Blockchain and Cryptocurrency, pp.1-4
2026 IEEE International Conference on Blockchain and Cryptocurrency (ICBC) (Brisbane, Australia, 01/06/2026–05/06/2026)
01/06/2026

Metrics

1 Record Views

Abstract

Blockchain Forensics Cross-Chain Security Graph Neural Networks (GNN) Risk Profiling Orbit Bridge
Cross-chain bridges have emerged as high-value targets, often serving as pseudo-anonymization layers that evade traditional behavioral heuristics. This paper presents TOPOS (TOpological Profiling of On-chain Subgraphs), a framework utilizing Heterogeneous Graph Networks (HGN) to identify malicious clusters through topological risk propagation. Evaluated against the 2024 Orbit Bridge exploit, TOPOS achieved a 100% detection rate of primary exploiters using only the bridge vault as a seed. While traditional heuristics assigned a mean risk score of only 0.134 to the attack cluster, TOPOS accurately identified 15 Critical and 28 High-risk entities with a mean score of 0.851. These results demonstrate that topological signals are superior to temporal behavior for characterizing the forensic blast radius of cross-chain exploits in real-time.

Details

Logo image