Over the past years, there has been a high increase in web sites using cloud computing. Like usual web site, those web applications can have most of the common web vulnerabilities, like SQL injection or cross-site scripting. Therefore, cloud computing has become more attractive to cyber criminals. Besides, in many cases it is necessary to comply with regulations like PCI DSS or standards like ISO/IEC 27001. To face those threats and requirements it is a common task to analyze web applications to detect and correct their vulnerabilities. The most used tools to analyze web applications are automatic scanners. But it is difficult to comparatively decide which scanner is best or at least is best suited to detect a particular vulnerability. To evaluate scanner capabilities some evaluation criteria have been defined. Often a web vulnerability classification is also used to evaluate scanners, but current web vulnerability classifications do not usually include all vulnerabilities. To face evaluation criteria which are not up-to-date and to have the fullest possible classification, in this paper a new method to map web vulnerability classifications is proposed. The result will be the vulnerabilities an automatic scanner has to detect. As classifications change over time, this new method could be executed when the existing classifications change or when new classifications are developed. The vulnerabilities described this way can also be seen as a web vulnerability classification that includes all vulnerabilities in the classifications taken into account.