Cyber Security News

New Obfuscation Trick Lets Attackers Evade Antivirus and EDR Tools

Researchers have unveiled a sophisticated new technique that allows attackers to bypass traditional Antivirus (AV) and Endpoint Detection and Response (EDR) solutions.

By exploiting how these defensive tools analyze command-line arguments—a core method of detecting suspicious activity—malicious actors can now cloak their intentions and evade detection with alarming efficiency.

The Power of Command-Line Obfuscation

As per Wietze’s report, Defensive security tools have long shifted from relying solely on identifying known malicious software to monitoring behavior and scrutinizing command-line arguments.

These arguments, supplied to applications at launch, often reveal whether an operation is benign or malicious. For example, terminating system processes or downloading files via system-native utilities using suspicious arguments typically sets off alarms.

However, attackers have adapted just as quickly. The newest trend, described by the creators of a tool called ArgFuscator, involves “command-line obfuscation”—a technique where the syntax of legitimate commands is manipulated to confuse security tools without altering the underlying behavior of the executable.

Unlike more familiar shell-based obfuscation (such as DOSfuscation or PowerShell obfuscation), this approach is independent of the shell environment and exploits parsing quirks in the executables themselves.

Screenshot showing DOSfuscation successfully obfuscating a command, but with the certutil execution ultimately showing up in unobfuscated form in ProcMon .Screenshot showing DOSfuscation successfully obfuscating a command, but with the certutil execution ultimately showing up in unobfuscated form in ProcMon .
Screenshot showing DOSfuscation successfully obfuscating a command, but with the certutil execution ultimately showing up in unobfuscated form in ProcMon .

How Obfuscation Techniques Work

ArgFuscator, an open-source project, documents dozens of obfuscation strategies that threat actors are now leveraging, including:

  • Option Character Substitution: Using unconventional characters (e.g., a hyphen instead of a slash) for command-line switches.
  • Character Substitution and Insertion: Swapping or adding Unicode characters to keywords (e.g., “reg eˣport” instead of “reg export”).
  • Quotes and Path Manipulation: Inserting superfluous quotes or unconventional paths to obscure the real command.
  • Value Transformations: Using numerical representations or odd formatting for values and addresses.

These tricks work on a wide array of trusted system executables (Living-off-the-Land binaries or LOLBINs), including commands like taskkill, reg, and curl.

Screenshot of the three described reg.exe obfuscation examples in action on a Windows 11 machine.

The result is that even well-configured security solutions may miss malicious activity if it arrives in a cloaked, yet technically valid, command-line format.

This development is especially concerning as “malwareless” attacks—intrusions that rely solely on built-in or trusted third-party tools—now account for the majority of observed breaches.

As attackers increasingly avoid dropping detectable malware in favor of misusing legitimate software, defenders face new hurdles.

The research behind ArgFuscator not only exposes these challenges but also provides defensive recommendations.

Security teams are urged to enhance detection rules by flagging unusual Unicode or excessive quoting, normalizing command lines before analysis, and correlating command activity with other indicators such as network traffic.

As attackers and defenders continue their high-stakes chess game, tools like ArgFuscator raise awareness, equipping security professionals with the knowledge—and warnings—they need to adapt for the next wave of cyber threats.

Find this News Interesting! Follow us on Google NewsLinkedIn, & X to Get Instant Updates!

Divya

Divya is a Senior Journalist at GBhackers covering Cyber Attacks, Threats, Breaches, Vulnerabilities and other happenings in the cyber world.

Recent Posts

Dutch Services Disrupted by DDoS Attacks From Russian-Affiliated Hacktivists

Multiple Dutch organizations have experienced significant service disruptions this week due to a series of…

3 hours ago

Seven Malicious Packages Exploit Gmail SMTP to Run Harmful Commands

A major supply chain security incident has rocked the Python open-source community as researchers at…

3 hours ago

CISA Issues New ICS Advisories Addressing Critical Vulnerabilities and Exploits

The Cybersecurity and Infrastructure Security Agency (CISA) has issued two new advisories revealing critical vulnerabilities…

5 hours ago

NVIDIA TensorRT-LLM Vulnerability Let Hackers Run Malicious Code

NVIDIA has issued an urgent security advisory after discovering a significant vulnerability (CVE-2025-23254) in its…

5 hours ago

CISA Issues Alert on Actively Exploited Apache HTTP Server Escape Vulnerability

The Cybersecurity and Infrastructure Security Agency (CISA) has issued an urgent alert regarding a newly…

7 hours ago

Disney Hacker Admits Guilt After Stealing 1.1TB of Internal Data

A 25-year-old man from Santa Clarita, California, has agreed to plead guilty to hacking into…

7 hours ago