Proofpoint says TA419 impersonated AI-policy figures and used a real-time Microsoft 365 phishing proxy against fewer than ten specialists. The campaign shows how attackers can capture authenticated sessions, while successful compromise and government attribution remain unproven.