Bug 2492456 (CVE-2026-44020)

Summary:	CVE-2026-44020 docling: Docling: Information disclosure via XML External Entity (XXE) vulnerability
Product:	[Other] Security Response	Reporter:	OSIDB Bzimport <bzimport>
Component:	vulnerability	Assignee:	Product Security <prodsec-ir-bot>
Status:	NEW ---	QA Contact:
Severity:	high	Docs Contact:
Priority:	high
Version:	unspecified	CC:	jkoehler, lphiri
Target Milestone:	---	Keywords:	Security
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	---
Doc Text:	A flaw was found in docling. An attacker could exploit an XML External Entity (XXE) vulnerability in the USPTO patent XML parser by crafting malicious XML files. This could allow the attacker to read arbitrary files from the server's filesystem, perform Server-Side Request Forgery (SSRF) attacks, or cause a denial of service by consuming system resources through entity expansion.	Story Points:	---
Clone Of:		Environment:
Last Closed:		Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description OSIDB Bzimport 2026-06-24 18:13:05 UTC

Docling simplifies document processing by parsing diverse formats and providing integrations with the generative AI ecosystem. From 2.13.0 until 2.74.0, the USPTO patent XML parser used the standard xml.sax.parseString() without protection against XML External Entity (XXE) attacks. An attacker could craft malicious USPTO patent XML files with external entity references that could read arbitrary files from the server filesystem, perform Server-Side Request Forgery (SSRF) attacks, or cause denial of service through entity expansion (Billion Laughs attack). The vulnerability affects three USPTO patent format parsers: ICE (v4.x), Grant v2.5, and Application v1.x. This vulnerability is fixed in 2.74.0.

Comment 2 Florencio Cano 2026-06-25 11:22:14 UTC

The vulnerability affects Docling versions >= 2.13.0 and < 2.74.0. Both tags ship docling==2.73.1, which falls in that range:

v3.4.0 (stable, built 2026-05-05) — docling 2.73.1 — AFFECTED
v3.5.0-ea.1 (early access, built 2026-05-28) — docling 2.73.1 — AFFECTED

The package is pulled as a pip dependency from the Red Hat internal PyPI mirror (console.redhat.com/api/pypi/public-rhai/rhoai/...) and is referenced in pipelines/training/autorag/documents_rag_optimization_pipeline/requirements.txt. Companion packages also present in both images: docling-core 2.70.2, docling-ibm-models 3.12.0, and docling-parse 4.7.3.

Comment 3 Florencio Cano 2026-06-25 11:33:27 UTC

Critical requires the flaw to lead to arbitrary code execution on the system. CVE-2026-44020 does not allow code execution. It is therefore excluded from Critical regardless of how easily it can be triggered. The Important definition covers flaws that "allow unauthenticated remote users to view resources that should otherwise be protected" and flaws that "allow remote users to cause a denial of service." This vulnerability satisfies both conditions simultaneously.