THREATSDAY Bulletin notes Kali Linux has integrated Anthropic’s Claude AI assistant via MCP to issue commands in natural language, alongside a broad run of related alerts and 15+ stories. The roundup covers a Belarusian Android spyware campaign, crypto phishing impersonations, rapid breakout times in 2025 with references to CrowdStrike and ReliaQuest, and Mac-focused ClickFix-style campaigns delivering the MacSync and related data-stealing tools.
It also highlights an ActiveMQ flaw (CVE-2023-46604) used to deploy LockBit ransomware, a Chrome crash-to-command trick with new variants, and widespread exposure to CVE-2025-8088 in WinRAR, reported by Stairwell. Additional items include AI-enabled audits of smart contracts, ongoing OpenAI–Paradigm work on EVMbench, and various spear-phishing and repo-squatting campaigns, all underscoring a trend of speed, deception, and blending into everyday activity.
The piece notes a Reuters report on Meta encrypting Messenger and Instagram despite internal warnings, and closes by urging vigilance against delayed patches and rushed clicks. 26 February 2026