Aggregated data from penetration tests and red team engagements suggests that many enterprise organizations are making progress in securing their networks against cyber adversaries.
External and internal assessments that pen-testing firms have conducted in recent years show that although organizational networks continue to present multiple weaknesses, attackers may be having a harder time finding and exploiting them from outside the network.
"I won't say that the days of 'point, click, and exploit' are over, but they sure are rare," says Chris Nickerson, CEO at pen-testing firm Lares. While security hardening, hygiene, patch management, password quality, and lack of visibility continue to remain big challenges, security organizations are evolving, he notes.
Increasingly, attackers are being forced to change their tactics and employ malware-less, "living-off-the-land" approaches to hide their malicious activity. "It is rare that 'exploitation' is the first hook into the environment anymore," Nickerson says. "Now tools and technology are required to observe normal system functions to determine if they are being used maliciously."
Lares recently analyzed data from hundreds of pen-test engagements to see what similarities it could find across enterprise networks. The results showed that accounts with weak and easily guessable passwords continue to be the biggest problem for most organizations. Other common vulnerabilities and attack vectors include weaknesses related to Kerberos authentication, excessive file system permissions, Window Management Interface (WMI)-enabled lateral movement, inadequate network segmentation, and improper access control.
Other pen-testing firms have found similar issues. In a report last year, Coalfire identified out-of-date software as the most commonly present threat in organizations where it conducted pen tests. Like Lares, Coalfire discovered that password flaws were a big problem. as was patch management, insecure protocols. and configuration protocols.
For security vendor Rapid7, meanwhile, the most common security issues that its pen testers encountered included weak transport layer security and vulnerabilities related to password management. In 72% of engagements, Rapid7 researchers were able to capture a user credential using generic password spraying, known defaults, or easily guessable passwords.
"Taken together, these information exposure issues do tend to be serious enough to report out to the client," says Tod Beardsley, director of research at Rapid7. "But penetration testers are rarely able to exercise these vulnerabilities to traverse the boundary between external and internal networks."
That's because in many instances, the network segmentation controls that organizations have implemented to separate internal and external networks appear to be working effectively, Rapid7 said in a pen test report of its own last year.
In externally based pen tests last year, Rapid7's pen testers were only able to gain internal LAN access just 21% of the time. In assessments of Web application security, Rapid7 testers were able to achieve total sitewide compromise just 3% of the time. The vendor found that the ongoing move to the cloud and the growing use of services like AWS, Microsoft Azure, and Google Cloud to host Web applications and data has made life significantly harder for attackers.
For internally focused engagements, Rapid7 found that most vulnerabilities and exploits were Windows-related simply because most internal environments are also Windows-based. Windows remote administration technologies like WMI and PsExec continue to provide opportunities for attackers to move laterally on breached networks. Rapid7 found that SMB relaying remains the most popular technique for attackers to gain an initial foothold. But like other pen-testing firms, Rapid7's own experience suggests that organizations are tightening their controls, Beardsley says.
"Enterprise security among the organizations that we pen test seems to be slowly but surely getting better," Breadsley says. "But we still struggle with the fundamentals: password management, network segmentation, and legacy systems."
One major issue that organizations still need to work through is what Lares' Nickerson describes as a "tooling debt." Over the years, many organizations have spent millions on security tools and have kept adding and churning through them. This has created a situation where organizations often have a large number of poorly integrated tools. "This is something we see at all layers of the program," Nickerson says.
Even with major technologies such as SIEM and SOAR platforms, some organizations can change them upward of three times in five years. "This turbulence in defense makes it quite hard for customization to stay in effect," Nickerson says. "Often, an alert or detection that was made from a previous test is lost" in the chaos.
According to pen-testing firms, the vulnerabilities and attack vectors they commonly encounter remain mostly the same regardless of the organization's size. But there are some differences. PowerShell-related exploits, for instance, are becoming less of a threat to the enterprise because larger organizations are adding more restrictions around its use, Beardsley says.
Similarly, while both large and small organizations have issues related to patch management, the reasons for the exposure are different. Small and midsize organizations often are reluctant to deploy patches because of concerns over "breaking" their system, says Nickerson.
"In the enterprise, it ends up being the sheer size of the environment," he adds. "There always seems to be a legacy box lurking in the corners, outside of the purview of the patch management program or asset inventory."