[TIKA-3074] Vulnerable "woodstox-core" is present inside Tika 1.23

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Assignee: Unassigned
• Reporter: Abhishek Chauhan
• Created: 17/Mar/20 14:25
• Updated: 18/Mar/20 18:18

Description

Short Description:  woodstox-core is a transitive dependency of Apache Tika. Checked the pom inside tika-app-1.23.jar, it seems that it is internally using 5.0.3 version of woodstox-core, which is vulnerable.

Root Cause : tika-app-1.23.jar; com/ctc/wstx/sax/WstxSAXParserFactory.class : [5.0.1 , 5.3.0]

Vulnerability: The woodstox-core package is vulnerable to Improper Restriction ofXML eXternal Entity [XXE] Reference. The setFeature and getFeature methods in WstxSAXParserFactory.class rely on the mSecureProcessing boolean value to be able to securely parse input XML. The boolean value, however, is set to false by default. Additionally, the class lacks support for properties XMLConstants.FEATURE_SECURE_PROCESSING and XMLInputFactory.IS_SUPPORTING_EXTERNAL_ENTITIES, which can make it possible for an attacker to conduct XXE attacks.

This vulnerability is addressed in the issue https://github.com/FasterXML/woodstox/issues/61 

Solution of the Vulnerability: Issue https://github.com/FasterXML/woodstox/issues/61 is fixed in version 5.3.0 of woodstox-core. Tika may need to upgrade the version of  this dependency, so consumers are not affected by transitive dependency.

 

Issue Links

• is fixed by: CXF-8245 Vulnerable "woodstox-core" is present inside Tika 1.23