Secure agentic AI for your Frontier Transformation

Microsoft Security Blog posts (Feb–Mar 2026) describe multiple high‑impact threats and defensive guidance: malicious AI browser extensions collected LLM chat histories and browsing data (≈900,000 installs, activity across >20,000 enterprise tenants); Tycoon2FA AiTM phishing‑as‑a‑service reached ~500,000 organizations monthly and was disrupted through industry law‑enforcement action; signed malware using a stolen EV certificate deployed legitimate RMM tools to establish persistent backdoors; OAuth redirection flows are being abused for phishing and malware delivery; and developer‑targeting via恶