A modern spin on the old-school hacker-behind-the-keyboard attack exceeded malware-borne ones worldwide last year, new incident report data from CrowdStrike shows.
Seasoned cybercriminals and nation-state attackers for some time now have been upping their game with new methods to mask their activities from security tools by blending in and posing as real users in the targeted organization's network - using stolen credentials and running legitimate tools to dig through victim systems and data, for instance. And for the first time in CrowdStrike's research and incident response engagement reporting, so-called "malware-free" attacks edged ahead of malware-based ones, at 51% to 49% in 2019. In 2018 and 2017, malware accounted for around 60% of all attacks globally, and malware-free attacks around 40%, according to CrowdStrike's data.
A malware-free attack in CrowdStrike's parlance is one where the method to gain entry into a victim organization doesn't employ a malicious file or file fragment to a computer disk. In addition to stolen credentials or legitimate tools, this type of attack also can execute code from memory and can only be detected with higher-level tools and techniques that spot unusual behavior, or via threat hunting.
Like the bad old days of hacking, much of the attack is driven by "hands-on keyboard" methods like command line interface, PowerShell, and hiding files and directories, according to CrowdStrike. "These techniques feature prominently in many sophisticated attacks, where a human adversary is engaged in the intrusion and is actively working toward an objective," according to the report.
Security experts worry that if attackers double down on malware-free attacks, security tools - and ultimately, targeted organizations - will be overwhelmed and unable to thwart them.
Michael Sentonas, CTO of CrowdStrike, says these malware-free attacks have gradually increased as attackers have found ways to bypass traditional security tools. But the attackers aren't stopping at commodity antivirus software. "Now they're starting to bypass next-generation AV products as well. That's driving a big escalation in malware-free attacks," he says. "If that hits 60% and above, it's going to be a big problem. "
The problem is that most organizations don't have the technology to discern between a legitimate user or an attacker who has stolen his or her credentials, Sentonas notes. "I want to track over the next 12 months the malware-free piece to see if that's a real interesting [longer term] trend there," he says.
Rapid7 also has seen attackers forgo malware when infiltrating their targets. "They are using valid credentials or reusing credentials from other breaches. That's hard [to defend against]," says Tod Beardsley, director of research at Rapid7. "It's not something you can easily automate defense on."
To help combat these threats, he says, organization needs to empower end users to be part of a security culture. "You have to build trust between IT security and the user base. In that vein you are building a culture of vigilance," he says, where if you see something that looks suspicious, there's a clear step on what to do about it, whether confirming it out-of-band or reporting it to the right people.
Most malware-free attacks last year occurred in North America, where three-fourths of attacks didn't deploy malware to get inside the victim organization, according to CrowdStrike. "It will be interesting as we go through this year: Will malware-free attacks continue to rise, and will that correlate in dwell time [of attackers]?" says Sentonas. "If we see that as a two- to four-year trend, we have a problem. There are more and more ways to evade security controls."
That requires a defense that accounts for the human element. "Now more attackers are human," says Chester Wisniewski, a principal research scientist with security vendor Sophos. "That's why setting tripwires" to detect legitimate tools being used for nefarious purposes is key, he says. For example, if the Nmap network monitoring tool is spotted running on a Web server in the DMZ, that should be a red flag, he says. "No one should be running it ... in the server in the DMZ," he says.
If a legitimate security tool is running at a time other than when it should be used, that constitutes an incident, he says. "You're not the only one using these tools," he says, noting that the bad guys are as well.
Check out The Edge, Dark Reading's new section for features, threat data, and in-depth perspectives. Today's top story: "With New SOL4Ce Lab, Purdue U. and DoE Set Sights on National Security."Kelly Jackson Higgins is the Executive Editor of Dark Reading. She is an award-winning veteran technology and business journalist with more than two decades of experience in reporting and editing for various publications, including Network Computing, Secure Enterprise ... View Full Bio