CVE-2025-66516: Apache Tika XXE via crafted XFA file in PDF

Automated analysis

Generated by VULONE's analysis model from the NVD record, CISA KEV and EPSS data on 18 September 2026. Confidence: high.

critical priorityCVSS 9.8 with network reachability, no authentication or interaction, and a very high EPSS score make this a top remediation priority despite no KEV listing.

What it is

Apache Tika's tika-core, tika-pdf-module and tika-parsers modules are vulnerable to XML External Entity injection when parsing a crafted XFA file embedded in a PDF. This CVE expands the scope of CVE-2025-54988: the flaw and its fix live in tika-core, so upgrading only the PDF module leaves users exposed, and 1.x releases carry the PDFParser in tika-parsers. It matters because Tika is widely embedded in ingestion pipelines, so untrusted documents can trigger the flaw without any user action.

Impact

An attacker can cause the parser to resolve external entities, leading to disclosure of local files or server-side request forgery, and the CVSS vector also rates high integrity and availability impact. In practice this gives an unauthenticated attacker a path to read data and disrupt the parsing service.

Attack surface

Reached remotely over the network by submitting a malicious PDF containing a crafted XFA file to any service that uses Tika for document parsing. No authentication or user interaction is required per the CVSS vector (AV:N/AC:L/PR:N/UI:N).

Exploitation

Not listed in CISA KEV and no public exploit reference is tagged, but EPSS is very high at 0.877 (99.75th percentile), indicating elevated near-term exploitation likelihood. Reference tags are only vendor and third-party advisories, so no confirmed in-the-wild exploitation is documented here.

What to do

• Upgrade tika-core to 3.2.2 or later; upgrading only tika-pdf-module is insufficient.
• For 1.x deployments, upgrade tika-parsers to a fixed release, since the PDFParser lives there in that line.
• Disable external entity and DTD processing in the XML parser configuration used for XFA and other untrusted XML.
• Isolate or sandbox Tika parsing services and restrict outbound network access to limit SSRF and file-read impact.
• Treat all uploaded PDFs as untrusted and validate or reject documents with XFA content where XFA is not required.

Detection

• Monitor Tika parsing logs for XML parser errors, entity resolution failures or unexpected external fetches during PDF processing.
• Alert on outbound network connections from Tika hosts to unusual destinations, which may indicate SSRF via external entities.
• Inspect uploaded PDFs for XFA streams and flag documents containing XFA for additional review.
• Track file access on Tika hosts for reads of sensitive local files outside expected document paths.

This assessment is produced automatically and is not human-reviewed. Verify against the vendor advisory before acting on it.