Severity by source
AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:N
Network-reachable but AC:H because a specific non-default route mapping untrusted data into the docling headers is required; C:H/I:H for out-of-directory read/write, A:N, no shell so no command execution.
Primary rating from Vendor (apache).
CVSS VectorVendor: apache
Lifecycle Timeline
6Blast Radius
ecosystem impact- 4 maven packages depend on org.apache.camel:camel-docling (2 direct, 2 indirect)
Ecosystem-wide dependent count for version 4.15.0.
DescriptionCVE.org
Improper Neutralization of Argument Delimiters in a Command ('Argument Injection') vulnerability in Apache Camel Docling component.
The camel-docling component invokes the external docling command-line tool by assembling an argument list in DoclingProducer and executing it through java.lang.ProcessBuilder. Custom CLI arguments supplied through the CamelDoclingCustomArguments exchange header (a List<String>) were appended to that argument list with insufficient validation: the original implementation relied on a denylist of disallowed flags and only rejected path values that contained a literal ../ sequence. As a result, a Camel route that forwards externally-influenced data into the CamelDoclingCustomArguments header (or into the path-bearing headers used to build the invocation) could cause the producer to pass unrecognized or unintended docling CLI flags to the subprocess, and could supply path-like argument values that resolved outside the intended directory through traversal sequences not caught by the literal ../ check. Because Camel itself builds the docling invocation from these values, the component is responsible for constraining them, and the weak validation allowed CLI-argument injection and directory traversal in the arguments passed to the external tool. The invocation uses the list-based form of ProcessBuilder, so a shell does not interpret the argument values; OS command injection through shell metacharacters was not possible, and the metacharacter rejection added by the fix is defense-in-depth. This issue affects Apache Camel: from 4.15.0 before 4.18.3.
Users are recommended to upgrade to a release that contains the CAMEL-23212 fix. On the mainline the fix is included from Apache Camel 4.19.0 (and later releases such as 4.20.0). For users on the 4.18.x LTS releases stream, upgrade to 4.18.3. The fix replaces the denylist with a strict allowlist of recognized docling CLI flags (rejecting any unrecognized flag, and rejecting producer-managed flags such as the output-directory flags), defensively rejects shell metacharacters in argument values, and normalizes path-like values with Path.normalize() before validating them so that traversal sequences which bypass a literal ../ check are detected. As defence in depth, route authors should avoid mapping untrusted message content into the CamelDoclingCustomArguments header and the path-bearing headers, and should strip Camel-internal headers from messages that arrive from untrusted producers.
Articles & Coverage 1
AnalysisAI
Argument injection and directory traversal in Apache Camel's camel-docling component (4.15.0 before 4.18.3) let attackers who can influence the CamelDoclingCustomArguments or path-bearing exchange headers inject unintended docling CLI flags and traversal-laden path values into the externally executed docling tool. Because the original DoclingProducer validation relied on a flag denylist and only rejected literal '../' sequences, crafted arguments could reach the subprocess and resolve files outside the intended directory, yielding high confidentiality and integrity impact but no OS command injection (ProcessBuilder uses the list form, so no shell interprets the values). There is no public exploit identified at time of analysis and the flaw is not in CISA KEV; EPSS is low (0.79%, 52nd percentile).
More in Apache Camel
View allRemote code execution in Apache Camel's camel-coap component (versions 4.14.0-4.14.5, 4.18.0 before 4.18.1, and 4.19.0)
Unsafe Java deserialization in Apache Camel's camel-infinispan component allows arbitrary code execution when the ProtoS
Header-injection in the Apache Camel camel-atmosphere-websocket producer lets an external sender hijack WebSocket messag
Header-filter bypass in the Apache Camel Undertow component (4.11.0-4.14.8, 4.15.0-4.18.3, and 4.19.0-4.21.x) lets remot
Remote code execution in Apache Camel 3.18.0-4.14.5 and 4.15.0-4.18.1 stems from CXF and Knative HeaderFilterStrategy im
Message header injection in Apache Camel's camel-mail component (3.0.0 before 4.14.6 and 4.15.0 before 4.18.1) lets an a
Remote code execution via unsafe Java deserialization affects the camel-pqc component of Apache Camel 4.18.0-4.18.2 and
Unsafe Java deserialization in Apache Camel's camel-consul component (ConsulRegistry / ConsulRegistryUtils.deserialize)
Cypher injection in Apache Camel's camel-neo4j producer allows attackers who control JSON key names in the CamelNeo4jMat
Remote code execution in the Apache Camel camel-hazelcast component allows an attacker who can join or reach the Hazelca
Remote code execution in Apache Camel's camel-vertx-http component (4.0.0-4.14.7, 4.15.0-4.18.2, 4.19.0) arises when a p
Blind out-of-band data exfiltration in Apache Camel 4.14.0-4.20.x arises because the default ObjectInputFilter pattern b
Same technique Path Traversal
View allVendor StatusVendor
Share
External POC / Exploit Code
Leaving vuln.today
EUVD-2026-41823
GHSA-rpv3-6645-2vqc