Document Processing Vulnerability in Docling Affects Generative AI Integrations
CVE-2026-44017

7.5HIGH

Key Information:

Status
Vendor
CVE Published:
24 June 2026

What is CVE-2026-44017?

Docling, a tool that streamlines document processing with diverse format parsing and integrations into the generative AI ecosystem, is affected by a vulnerability in versions prior to 2.91.0. This flaw arises from the EasyOCR model download functionality, which improperly extracts ZIP archives without proper path validation. This oversight may allow attackers to exploit the model download source through various means such as supply chain attacks or DNS spoofing. Consequently, they can potentially write files to unintended locations within the system, which could lead to severe consequences, including remote code execution through overwritten files, installation of persistent backdoors by altering startup scripts or SSH keys, and critical data corruption or system compromise. The issue has been addressed in version 2.91.0.

Affected Version(s)

docling < 2.91.0

References

CVSS V3.1

Score:
7.5
Severity:
HIGH
Confidentiality:
High
Integrity:
High
Availability:
High
Attack Vector:
Network
Attack Complexity:
High
Privileges Required:
None
User Interaction:
Required
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.