MITRE Names 2019’s Most Dangerous Software Errors
Eight years ago, a list of the world’s most dangerous software errors was published by problem-solving nonprofit the MITRE Corporation. Yesterday saw the long-awaited release of an updated version of this rag-tag grouping of cyber-crime’s most wanted.
The Common Weakness Enumeration (CWE) Top 25 Most Dangerous Software Errors list (CWE Top 25) is a roundup of the most widespread and critical weaknesses that can lead to serious vulnerabilities in software.
What makes these bad boys so lethal is that they are often easy to find and exploit. And once attackers have gotten their grappling hooks into the errors, they are frequently able to completely take over execution of software, steal data, or prevent the software from working.
Each error was given a threat score to communicate its level of prevalence and the danger it presents. Topping the table of treachery with a threat score of 75.56 and leading by a huge margin is “improper restriction of operations within the bounds of a memory buffer.”
The second-most lethal error was determined to be “improper neutralization of input during web page generation,” also known as cross-site scripting, which had a threat score of 45.69.
In 2011, a subjective approach based on interviews and surveys of industry experts was used to create the list. In 2019, the list’s compilers took a data-driven approach, leveraging National Vulnerability Database (NVD) data from the years 2017 and 2018, which consisted of approximately 25,000 CVEs.
MITRE’s goal is to release an updated list each year based on data from that specific year. Asked why the gap between the first two lists was so long, a MITRE spokesperson answered: “Based on the previous methodology employed for the 2011 CWE Top 25 List, it was clear that there was no basis upon which to credibly change the list.
“As new methodologies were explored, and upon selection of the current data-driven approach, it became valuable to produce a new list because it would validate whether or not the new data-driven methodology would result in a different list. And, since it did result in a different list, community stakeholders now have a new list to consume that is evidence-based and different from the 2011 list.”
The lists are indeed different, but both include some of the same offenders. Explaining why, the spokesperson said: “Significant work remains in the community to educate developers, improve analysis tools, and for consumers of software products to understand that weaknesses exist, and that they have the ultimate leverage with respect to evaluating products and selecting those products that deliberately work weaknesses out.
“Effective security can exist only if a broad number of stakeholders demand that it does. The 2019 CWE Top 25 List is a tool that different stakeholders can use to understand what the most prevalent weaknesses are and how to orient themselves toward defending against them.”