Phishing Campaign Uses ASCII Smuggling to Evade Email Filters

A phishing campaign has been using ASCII smuggling to evade email filters and target finance-themed emails. The technique, popularized in AI prompt injection research, involves hiding instructions inside invisible Unicode tag characters. Microsoft researchers observed a high-volume phishing campaign using this method, with a sharp increase in hits on February 9, 2026.

The attackers used the U+E0000 to U+E007F range of Unicode tags to split financial lure words, preventing email filters from parsing them. This technique is similar to AI prompt injection, where attackers hide instructions inside invisible tag characters to induce an AI assistant to follow threat actor-controlled instructions.

Microsoft Defender for Office 365 telemetry showed that the majority of messages were flagged by layered protections rather than a single Unicode-specific signal. The attack was observed to have a strict weekly cadence, with high volumes on weekdays and almost complete silence on weekends.