Security

ASCII Smuggling Phishing Evasion Hits Microsoft 365

3 min read

Summary

Microsoft researchers observed a high-volume phishing campaign using invisible Unicode tag characters, known as ASCII smuggling, to evade email filters in finance-themed lures. The technique, previously associated with AI prompt injection, shows how AI-era evasion methods are now crossing into traditional phishing, making layered email security and hunting more important for defenders.

Need help with Security?Talk to an Expert

Introduction

Microsoft has uncovered a notable shift in attacker behavior: a technique made popular in AI prompt injection research is now being used in mainstream phishing campaigns. For Microsoft 365 and security teams, this matters because it shows how evasions developed for AI systems can quickly become practical tools for email-based attacks.

What is new

Microsoft Security Research observed a large phishing campaign using invisible Unicode tag characters from the U+E0000 to U+E007F range. This method, called ASCII smuggling, lets attackers split or hide lure words such as “funding” so that the text appears normal to users but may be parsed differently by detection systems.

Key findings include:

  • The activity was first surfaced through Microsoft Defender for Office 365 prompt injection protection research.
  • Signature hits spiked sharply on February 9, 2026.
  • Volume jumped from about 21,000 messages on February 8 to more than 1.3 million the next day.
  • Peak volume exceeded 2.3 million messages on February 11.
  • The campaign largely followed a weekday-only sending pattern for roughly three months.
  • Microsoft linked the activity to a broader finance-themed phishing campaign using around 150 sender domains.

Why this matters for defenders

This finding highlights an important security trend: techniques first discussed in AI red-teaming and prompt injection are now being repurposed for classic phishing evasion. In this case, the same invisible Unicode characters that can hide instructions from humans while exposing them to AI models were used to obscure phishing keywords from email filters.

Microsoft also noted that detection did not rely on a single Unicode-specific rule alone. Instead, layered protections in Defender for Office 365 flagged most messages, which reinforces the value of defense in depth rather than depending on one signature.

Impact on IT administrators

For administrators, this means email hunting and detection logic should account for unusual Unicode usage, especially the Unicode Tags block. However, tuning is essential because legitimate content can also trigger detections, such as subdivision flag emojis for England, Scotland, and Wales.

Security teams should expect more crossover between AI-related attack research and traditional threat activity. Techniques that once seemed niche may quickly appear in large-scale commodity phishing.

Next steps

  • Review Defender for Office 365 detections and hunting queries for suspicious Unicode tag usage.
  • Validate that layered phishing protections are enabled and monitored.
  • Train analysts to recognize invisible-character obfuscation as a phishing signal.
  • Watch for finance-themed lure campaigns and unusual weekday-volume patterns.
  • Revisit email security controls to ensure they normalize or inspect hidden Unicode content where possible.

As attackers adapt AI-era evasion methods for email, defenders should update hunting strategies now rather than wait for these techniques to become standard practice.

Need help with Security?

Our experts can help you implement and optimize your Microsoft solutions.

Talk to an Expert

Stay updated on Microsoft technologies

ASCII smugglingphishingMicrosoft Defender for Office 365Unicodeemail security

Related Posts

Security

Microsoft Sentinel Named a 2026 SIEM Leader

Microsoft has been named a Leader in the 2026 Gartner Magic Quadrant for SIEM, highlighting the continued momentum behind Microsoft Sentinel. The announcement also underscores Microsoft's push toward integrated, AI-powered security operations through deep Sentinel integration with the new Integrated Security Operations Center in Microsoft Defender.

Security

Microsoft Perception GA: Agentic Security at Ignite

Microsoft has made Microsoft Perception generally available, introducing an agentic security platform that coordinates red, blue, and green team agents through automated playbooks. Announced at Ignite 2026, the update also expands Defender and Sentinel capabilities to help security teams detect, investigate, and remediate AI-speed threats while keeping administrators in control.

Security

Device Code Phishing Risks for AI Coding Agents

Microsoft warns that device code phishing can be used against AI coding agents such as GitHub Copilot CLI, not just user mailboxes or cloud apps. In a controlled lab, a single approved device code sign-in could enable remote session steering, memory poisoning, synced session exposure, and private repository access depending on account permissions and enabled features.

Security

Microsoft 365 Outage Phishing Campaign Targets Users

Microsoft Threat Intelligence reported a high-volume phishing campaign that impersonates Microsoft 365 service-health notifications to trick users into clicking malicious links. The campaign matters because it uses outage-themed urgency and recipient-specific tracking to target multiple industries, increasing the risk of credential theft and follow-on account compromise.

Security

SCKit ScreenConnect Attack: MSPs Must Hunt Now

Microsoft Threat Intelligence confirmed a campaign abusing ConnectWise ScreenConnect to deploy unauthorized remote-access clients and the SCKit worm on MSP-managed endpoints. The activity can also target technician sessions, making this a high-priority incident for defenders because removing a rogue client alone does not eliminate persistence or invalidate exposed sessions.

Security

AI Agent Supply Chain Attacks Target Coding Tools

Microsoft Security researchers detailed two post-training supply chain attack paths that can backdoor coding and terminal agents: dataset poisoning and model poisoning. The findings matter because compromised agents can misuse tools, execute commands, and access sensitive environments even when prompts appear benign, reinforcing the need for stronger controls around training data, model provenance, and tool permissions.