Thoughts on the Latest Apache Struts Vulnerability CVE-2018-11776 operates at a far deeper level within the code than all prior Struts vulnerabilities. This requires a greater understanding of the Struts code itself as well as the various libraries used by Struts.
About a week ago, a security researcher disclosed a critical remote code execution vulnerability in the Apache Struts web application framework that could allow remote attackers to run malicious code on the affected servers.
The vulnerability (CVE-2018-11776) affects all supported versions of Struts 2 and was patched by the Apache Software Foundation on August 22. Users of Struts 2.3 should upgrade to 2.3.35; users of Struts 2.5 need to upgrade to 2.5.17. They should do so as soon as possible, given that bad actors are already working on exploits.
"On the whole, this is more critical than the highly critical Struts RCE vulnerability that the Semmle Security Research Team discovered and announced last September," Man Yue Mo, the researcher who uncovered the flaw, told the media, referring to the vulnerability (CVE-2017-9805) that hackers used to compromise Equifax last year, which led to the lifting of personal details of over 148 million consumers.
More Critical than Equifax Vulnerability
Struts, an open source framework for developing web applications, is widely used by enterprises worldwide, including many Fortune 100 companies. In 2017, the Equifax credit reporting agency used Struts in an online portal, and due to Equifax not identifying and patching a vulnerable version of Struts, attackers were able to capture personal consumer information such as names, Social Security numbers, birth dates and addresses of over 148 million US consumers, nearly 700,000 UK residents, and more than 19,000 Canadian customers.
Over the past year, I've lost track of the number of people asking whether they should migrate from Apache Struts to some other framework. Behind all those requests was an implicit fear that more critical issues were present.
Unfortunately, changing application frameworks isn't as easy as adopting a new pizza chain or even buying a new car. Rather its more akin to dumping your favorite sports team for another. The investment your development team made in understanding the framework and implementing features supported by that framework likely won't transfer to an alternate option — assuming a compatible one even exists.
Instead, development teams need to accept that frameworks will have issues and that those issues are likely the result of attempts by the framework developers to appeal to the broadest audience. This means that you shouldn't expect the framework to magically protect against security issues you can address in your code. Input validation is a perfect example of this — bad data going in often results in bad data coming out, or in operations that we never anticipated would be performed.
Modern software is increasingly complex, and identifying how data passes through it should be a priority for all software development teams. To give some background, developers commonly use libraries of code, or development paradigms which have proven efficient, when creating new applications or features. This attribute is a positive when the library or paradigm is of high quality, but when a security defect is uncovered this same attribute often leads to a pattern of security issues.
In the case of CVE-2018-11776, the root cause was a lack of input validation on the URL passed to the Struts framework. In both 2016 and 2017, the Apache Struts community disclosed a series of remote code execution vulnerabilities. These vulnerabilities all related to the improper handling of unvalidated data. However, unlike CVE-2018-11776, the prior vulnerabilities were all in code within a single functional area of the Struts code. This meant that developers familiar with that functional area could quickly identify and resolve issues without introducing new functional behaviors.
CVE-2018-11776, on the other hand, operates at a far deeper level within the code, which in turns requires a deeper understanding of not only the Struts code itself but the various libraries used by Struts. It is this level of understanding that is of greatest concern — and this concern relates to any library framework. Validating the input to a function requires a clear definition of what is acceptable.
Equally critical are requirements that the documentation for public functions clearly document how those function use and operate on any data passed to them. This forms a contract between the framework and the consumer, and sets expectations surround proper usage. Absent this contract and associated documentation, it's difficult to determine if the code is operating correctly or not. This contract becomes critical when patches to libraries are issued because it is unrealistic to assume that all patches are free from behavioral changes.
Shortly after the Apache Software Foundation released its patch, a proof-of-concept (PoC) exploit of the vulnerability was posted on GitHub. The PoC included a Python script that allows for easy exploitation. The firm that discovered the PoC, threat intelligence company Recorded Future, also said that it has spotted chatter on underground forums revolving around of the flaw's exploitation. Companies that don't want to become the next Equifax should immediately identify where and what version of Apache Struts they have in use and apply the patch as needed.
Black Hat Europe returns to London, Dec. 3-6, 2018, with hands-on technical Trainings, cutting-edge Briefings, Arsenal open-source tool demonstrations, top-tier security solutions and service providers in the Business Hall. Click for information on the conference and to register.
Tim Mackey is Principal Security Strategist, CyRC, at Synopsys. Within this role, he engages with various technical communities to understand how to best solve application security problems. He specializes in container security, virtualization, cloud technologies, distributed ... View Full Bio