A hacker used Claude Code on Korean banks. The logs held a CV request.
CrowdStrike researchers found exposed AI session logs from attacks on at least five South Korean lenders, The Register and Reuters reported on 8 October.
Shinhan Bank said information on about 25,000 customers was compromised; The Register put KB Kookmin Bank's count at 119.
According to CrowdStrike, the attacker worked with Anthropic's Claude Code, an AI coding assistant, and in separate sessions with other AI models, alongside ARTEX, a Chinese-developed open-source penetration-testing tool of the kind security teams use to test their own defences. In the logs, one prompt asked for a CV for a person listing a university in Guangdong, China, with an age that does not match the birth date given.
CrowdStrike said the details likely belong to the attacker but "it cannot definitively establish that connection." It has not attributed the attacks to a named group. Anthropic did not immediately respond to requests for comment, Reuters reported.
The AI sessions used in the attack also left a record of it, and the record surfaced through a security firm's research. Korean lawmakers will question bank executives at a hearing on 19 October.
Who in your company can see what staff type into AI tools?
Image: AI-generated.
Sources
Every story here is open to read. The ERP LEADERS brief goes one step further.
The week in enterprise software, in one email: the stories that mattered, the failures with their figures, one case taken apart. Read a sample or sign up for the brief.
Welcome back. · Issue 01
