Apache MyFaces 1.1.7 and 1.2.8, as used in IBM WebSphere Application Server and other applications, does not properly handle an unencrypted view state, which allows remote attackers to conduct cross-site scripting (XSS) attacks or execute arbitrary Expression Language (EL) statements via vectors that involve modifying the serialized view object.
Information disclosure vulnerability in Apache MyFaces Core 2.0.1 through 2.0.10 and 2.1.0 through 2.1.4 allows remote attackers to inject EL expressions via crafted parameters.
In the default configuration, Apache MyFaces Core versions 2.2.0 to 2.2.13, 2.3.0 to 2.3.7, 2.3-next-M1 to 2.3-next-M4, and 3.0.0-RC1 use cryptographically weak implicit and explicit cross-site request forgery (CSRF) tokens. Due to that limitation, it is possible (although difficult) for an attacker to calculate a future CSRF token value and to use that value to trick a user into executing unwanted actions on an application.