TechNadu • 3h
Kaspersky Compromise Assessments and Anthropic Claude Code Data Leakage
Kaspersky's 2025 Compromise Assessments reveal systemic detection failures, with threat actor dwell times reaching four years and 52% of high-severity breaches persisting beyond 90 days. Concurrent research into Anthropic's Claude Code highlights a critical data exfiltration vector where the tool captures confidential files and session data, leading to corporate bans by organizations like Alibaba. Technical artifacts include the NSABuffMiner crypto-mining malware and AI-driven exfiltration mechanisms, indicating a dual threat of long-term persistent compromises and emergent AI-driven supply chain leakage across government (29%) and financial (17%) sectors.
Links:TechNadu, Kaspersky Securelist, Reddit, Youtube, Zscaler, Beankinney, Coder, Bitdefender, Scmp, Letsdatascience, Content •