Wallet custody, cryptographic trust boundaries, vulnerabilities, audits and practical defensive habits

Bitget Wallet’s phishing alerts rose sharply, but flagged requests are not confirmed losses

Before signing a wallet transaction, the key question is what permissions it would grant. In a report published November 11, 2026, at 18:00 UTC+08:00, Bitget Wallet said it was expanding AI-assisted checks intended to identify phishing risks before users sign.

According to chief operating officer Alvin Kan, the wallet flagged about 37,000 signing requests per week for potential phishing in the third quarter of 2026, up from 5,000 per week in the second quarter—a 7.4-fold increase. Kan said these figures count flagged signing interactions, not unique attacks or confirmed financial losses. He told Cointelegraph that the rise likely reflected both increased threat activity and improved detection, but the data could not establish how much each contributed. Bitget Wallet introduced AI-assisted risk detection in Q2, he said.

Bitget Wallet said its expanded checks examine website code and domain registration details for signs of fraud and simulate transactions to identify hidden permissions before users sign. It also said AI helped identify 90% of newly blocked risky domains in October, compared with 36% in Q2. Nearly 70% of domains flagged by its AI that October had no matching alerts in the external threat feeds it checked.

The report cited other phishing threats: a trader lost nearly $1 million in July after signing a malicious Ethereum token approval, while an August Rapid7 investigation uncovered a campaign targeting 885,000 phone numbers with links to sites impersonating crypto wallet providers. For holders, the practical distinction is to inspect what a signing request authorizes rather than treat a wallet alert—or the absence of one—as proof of an outcome.

Original report: Cointelegraph

Further reading