TasteScore TrustGraph research case study
Binance hot-wallet compromise assessment
A historical research retrospective examining approval activity, a newly deployed sweeper, and collection-wallet staging in a proposed Binance hot-wallet compromise sequence.
Pre-sweep risk signals
The case study compares a frozen pre-sweep view with a post-sweep view. It examines an unlimited spender approval, an unverified contract, a shared spender associated with Binance 14 and Binance 16, and a staged collection address.
The experimental pre-sweep policy assigns a risk-adjusted score of 205 and a REVIEW decision. The post-sweep view assigns a score of 0 and a BLOCK decision. These are retrospective research outputs, not proof that the incident was predicted.
Method and limitations
The page explains its proposed scoring weights and links to the referenced Ethereum addresses on Etherscan. The scoring hypothesis is not calibrated as a probability of compromise, is not a live detection, and requires blinded validation against benign activity and additional incidents.
Its purpose is to examine how graph-aware controls might review privileged wallet approvals before value moves.