Apache Tika is a widely used toolkit for content analysis and extraction, often integrated into backend systems for data ingestion pipelines, Content Management Systems (CMS), and search indexing services. Its ability to process hundreds of file types makes it a critical component in high-volume data environments.

The impact of this vulnerability is Critical. An unauthenticated attacker can achieve Server-Side Request Forgery (SSRF) and local file exfiltration simply by submitting a malicious PDF file. This allows attackers to steal sensitive configuration files, credentials, and proprietary data directly from the server’s filesystem. Furthermore, the SSRF capability allows attackers to perform lateral movement and scan the internal network, accessing and potentially compromising other internal services reachable by the Tika server.

This vulnerability poses a significant risk to organizations with publicly exposed file upload features that rely on Apache Tika for processing. Given that public exploit code is available, the likelihood of active exploitation is high. Internet-facing applications using vulnerable versions of Tika are immediately at risk of data breaches and internal network compromise.

Technical Summary

The root cause of this vulnerability is CWE-611: Improper Restriction of XML External Entity References, commonly known as XXE injection. The flaw is specifically present within the tika-parser-pdf-module.

When Apache Tika parses a PDF file, it attempts to process various components, including XML-based Form Architecture (XFA) data if present. The vulnerability is triggered because the underlying XML parser used for the XFA stream is not securely configured to disable the resolution of external entities.

The attack chain is as follows:

1. An attacker creates a PDF document containing a malicious XFA payload. This payload defines an external entity pointing to a local file URI (e.g., file:///etc/passwd) or an internal network resource.
2. The malicious PDF is uploaded to an application that uses a vulnerable version of Apache Tika for parsing.
3. The tika-parser-pdf-module processes the PDF, identifies the XFA data, and passes it to the misconfigured XML parser.
4. The parser resolves the malicious external entity, causing the server to read the content of the specified file or execute a request to an internal network endpoint.
5. The exfiltrated data is embedded into the extracted text content, which can then be returned to the attacker or stored in an accessible location.

A conceptual example of a malicious XXE entity within XFA data is as follows:

<!DOCTYPE xfa [
  <!ENTITY xxe SYSTEM "file:///etc/shadow">
]>
<form>
  <textfield>...&xxe;...</textfield>
</form>

Vulnerable Versions: Apache Tika versions from 1.13 to 3.2.1 are vulnerable.
Fix Availability: A fix is expected in a future release. Users should monitor the official Apache Tika project for patch announcements.

Recommendations

• Apply the patch immediately: Update all instances of Apache Tika to the latest stable version as soon as a patch is released by the Apache Software Foundation. Monitor the official project page for security bulletins.

• Mitigations:

  • If patching immediately is not possible, consider temporarily disabling the tika-parser-pdf-module if PDF processing is not a critical requirement.
  • Implement strict network egress filtering on servers running Apache Tika to block outbound connections to unexpected internal and external IP addresses, limiting the impact of the SSRF vector.
  • Apply a secure XML parsing configuration globally, if possible, ensuring that all XML parsers disable external entity and DTD processing by default.

• Hunting and Monitoring:

  • Analyze application logs for PDF parsing errors that contain snippets of content from server-side files or responses from internal network services.
  • Monitor server and network logs for unusual outbound network traffic originating from Tika application servers, particularly requests to internal metadata services (e.g., 169.254.169.254) or other sensitive endpoints.
  • Inspect file upload logs for submitted PDFs that trigger exceptions related to XML or XFA parsing.

• Incident Response:

  • If a compromise is suspected, immediately isolate the affected server from the network to prevent further data exfiltration or internal lateral movement.
  • Preserve server logs, application logs, and any suspicious PDF files for forensic analysis.
  • Assume that any secrets, credentials, or API keys stored on the server’s filesystem have been compromised and initiate rotation procedures.

• Defense in Depth:

  • Run applications that use Apache Tika in isolated environments or containers with minimal filesystem and network access.
  • Apply the principle of least privilege to the service account running the Tika process, ensuring it does not have read access to sensitive system files.
  • Implement robust network segmentation to prevent servers processing untrusted data from establishing connections with critical internal infrastructure.