CVE-2025-47177 is not the instrument here. The instrument is an SMS spoofing chain that the FBI documented in a June 2026 advisory: Russia-linked actors sent texts impersonating mobile carriers, routed targets to credential-harvesting pages, and extracted session tokens for Signal and WhatsApp at scale across at least three NATO-adjacent countries. The advisory names no specific GRU or FSB unit, which is the correct call given what the artifacts support. The claim that these were "support texts" is technically accurate and almost entirely misleading, in the same way that describing a car-bomb as "a vehicle" is technically accurate. The harvested tokens enable persistent message access without triggering re-authentication -- meaning the access survives a password reset, which is the part that matters to any FSI compliance officer reading MAS TRM Notice 655's incident-reporting obligations.

The pressure point for APAC institutions is not the SMS vector, which is old. It is the convergence the week's other news surfaces: prompt injection is now documented as operationally active against enterprise RAG pipelines and agentic model routers (MITRE ATT&CK technique T1055.015 is the closest mapping, though the tactic is still ahead of the taxonomy), and Alibaba's Qwen-3 and DeepSeek-V3 have closed enough of the capability gap with US models that the assumption of Western-vendor AI-security superiority -- the assumption quietly embedded in every AI security audit checklist written before Q1 2026 -- is no longer empirically defensible. Annual audit cadences were calibrated to a threat environment that no longer exists. The specific signal to watch: whether MAS revises Notice 655's 14-day major incident reporting window before the next APAC FSI tabletop season, which is typically scheduled for October.