TA419 Deploys Frameless BitB AitM Against U.S. AI Policy Targets Since April 2025

TA419 opened campaigns with innocuous subject-matter invites impersonating economists, OSTP alumni, and an Anthropic employee before escalating to shortened URLs that passed Cloudflare Turnstile and delivered adversary-in-the-middle Microsoft login pages. The Frameless BitB technique replaced iframe methods with HTML/CSS/JS substitutions to capture session cookies without visible failure. Proofpoint telemetry showed successful credential relay to real Microsoft infrastructure, allowing undetected session hijacking.

Procurement records and prior TA419 activity since April 2025 reveal consistent focus on U.S. and Japan defense, energy, and foreign-policy nodes. The AI-policy pivot aligns with documented Chinese intelligence priorities around export controls and model-distillation disputes rather than a new operational line. Official attribution rests solely on Proofpoint infrastructure clustering; no independent packet or endpoint artifacts have been released for third-party verification.

The campaigns expose gaps in think-tank and university identity controls that federal grant and clearance processes do not audit. Passkey mandates remain absent from most non-profit and academic Microsoft tenants. Continued expansion of the same redirect-and-proxy stack against additional Japan-linked targets is the pattern most consistent with existing contract and targeting data.