Filter Evasion: Phishing Campaign Hides Invisible Characters in Email Subject Lines
A newly uncovered phishing campaign, identified by researchers at the Internet Storm Center, showcases a remarkably unconventional method of evading email filters—by embedding invisible characters within message headers. Specifically, the attackers employ soft hyphen characters inserted into subject lines through MIME encoding, creating messages that appear perfectly normal to recipients while concealing structural anomalies that hinder automated inspection.
This technique exploits the peculiarities of RFC 2047 encoding, a standard used for transmitting non-ASCII characters in email headers. In the intercepted campaign, the message subject was divided into two segments, each forming a Base64-encoded MIME block containing UTF-8 characters. Upon decoding, researchers discovered several soft hyphen (U+00AD) characters—also known as the HTML entity ­. These symbols remain invisible in most email clients, including Outlook, yet persist within the underlying data of the message.
The advantage for attackers lies in their ability to fragment keywords with unseen characters, rendering them difficult for signature-based filters to recognize. While the text appears coherent to human readers, content-based detection systems interpret it as disjointed fragments that fail to match any known attack patterns.
What makes this case particularly noteworthy is the intentional insertion of invisible symbols within the subject line, a technique rarely observed outside message bodies. Microsoft experts had already documented similar injections in 2021 within email content and headers, but such practices remain infrequent—leaving a gap in many existing defenses. Current filtering still focuses primarily on message bodies, links, and attachments, while headers often receive only superficial scrutiny.
In addition to the manipulated subject lines, the attackers inserted soft hyphens into the email body, subtly breaking word structures and further obscuring the content. Victims were redirected to a spoofed web interface designed to mimic a legitimate email login page, with the ultimate goal of stealing user credentials.
To mitigate such threats, security experts recommend updating email filtering policies to include comprehensive analysis of all message components for hidden characters. Administrators should also inspect MIME headers for anomalous encodings and adopt advanced analytical mechanisms that go beyond traditional keyword and signature detection.
Support Our Threat Intelligence
If you find our technology report and cybersecurity news helpful, consider supporting our work.