AI Weekly Malaysia

This ThreatsDay roundup argues the week's attacks came from ordinary-looking operations that do more than expected: a model inspection step that can execute code, a cache that can mix up requests, and public secrets that stay usable for years. The concrete items given are OFAC sanctioning 10 targets tied to a Tren de Aragua ATM jackpotting scheme using Ploutus malware, with $40.73 million in reported losses across more than 1,500 U.S. attacks, and roughly $6.1 million in inflows to seven designated crypto wallets since March 2022. It also notes EtherHiding, where actors hide malware instructions on public blockchains so they cannot easily be seized or taken down, plus a claim of 543K live secrets and a model-inspection RCE that the excerpt does not name or detail.

Why: Two of the listed items sit directly in AI and dev workflows: 'a model check can run code' means loading or inspecting third-party model artifacts is a code-execution decision, not a read-only one, and 543K live secrets implies leaked keys stay valid long after the leak. The excerpt does not name the affected tool, CVE, or vendor, so you cannot patch from this alone - treat it as a prompt to check whether your model-loading path uses safe formats and whether your own repos are still leaking usable credentials. The ATM jackpotting and sanctions items have no practical bearing on most builders in this audience.

Tom's Hardware reports on a Chainalysis report finding blockchain-assisted cyberattacks up more than fivefold since last year, driven mainly by North Korean and Iranian state actors and Russian-speaking criminal groups. The technique, called Blockchain Dead Drops (BDD), stores malicious payloads in on-chain transactions and smart contracts so infected devices can retrieve them on demand from public, censorship-immune blockchains rather than from servers that can be seized or blocked. The excerpt also claims open-weight LLMs are linked to an increase in attacks, but cuts off mid-sentence before any supporting detail.

Why: The concrete operational change named here is payload hosting moving off takedown-able servers onto public chains, which means incident response that relies on blocking domains, IPs, or seizing C2 infrastructure may not remove the payload source. If your detection stack only watches HTTP/DNS egress, BDD retrieval traffic is a different channel you have not instrumented. Note that the open-weight LLM claim is asserted in the headline but the excerpt ends before showing the evidence, so do not repeat it as fact.