Fix Apache Tika XXE CVE-2025-66516

[cursor]

Summary

• Customer security teams escalated CVE-2025-66516 (Apache Tika XFA XXE/SSRF) that exposes file read and SSRF vectors in our bundled Tika plugin.
• dotCMS currently ships Apache Tika 2.8.0 via independent-projects/core-plugins/tika-plugin/pom.xml and 1.28.5 in osgi-base/system-bundles/pom.xml, both of which are vulnerable.
• We must upgrade every Tika artifact we publish (core plugin, shaded bundles, docker images) to the patched Tika release and communicate availability to customers.

Details

• CVE: https://nvd.nist.gov/vuln/detail/CVE-2025-66516 (CVSS 10.0)
• Current plugin property: <tika.version>2.8.0</tika.version>
  • independent-projects/core-plugins/tika-plugin/pom.xml
  • osgi-base/system-bundles/pom.xml still references 1.28.5 when embedding the plugin.
• Vulnerability allows malicious PDFs with XFA forms to exfiltrate local files and perform SSRF once parsed. Public PoC already exists.

Acceptance Criteria

• Upgrade all Apache Tika dependencies (core + parsers) to the first fixed version (latest 2.x containing the XXE patch; confirm upstream release notes).
• Ensure OSGi/system bundle packaging embeds the updated plugin and no longer distributes prior versions.
• Rebuild/publish patched dotCMS containers or plugins and update dev.dotcms.com/docs/known-security-issues with mitigation guidance.
• Add regression coverage (unit or integration) around PDF ingestion to ensure no parsing regressions with the new Tika.
• Communicate availability to Support for responding to ticket 34718.