GHSA-2WPJ-35WR-VCH7
Vulnerability from github – Published: 2026-09-28 12:31 – Updated: 2026-09-28 15:31Apache Karaf's instance-management service (InstanceServiceImpl) builds the command line used to launch a child Karaf JVM by string concatenation, then executes it through /bin/sh (Unix) or cscript (Windows). The caller-supplied javaOpts value is spliced into that string unquoted. A javaOpts value containing shell metacharacters (;, |, `, $(...)) is interpreted by the shell instead of being passed to the JVM as an option, giving arbitrary OS command execution as the Karaf process user.
Reachable via the shell commands instance:create, instance:start, instance:restart, instance:change-opts, and the equivalent InstanceMBean JMX operations (createInstance, startInstance, changeJavaOpts, cloneInstance).
Mitigation * Set karaf.secured.command.compulsory.roles=admin in etc/system.properties to close the fail-open gap for all unconfigured command scopes. * Restrict which principals can reach instance:* commands and InstancesMBean via etc/users.properties role assignments. * Treat javaOpts passed to instance:create/instance:start/instance:change-opts/InstancesMBean as untrusted input only from fully-trusted operators.
{
"affected": [],
"aliases": [
"CVE-2026-91006"
],
"database_specific": {
"cwe_ids": [
"CWE-78"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-09-28T11:16:48Z",
"severity": null
},
"details": "Apache Karaf\u0027s instance-management service (InstanceServiceImpl) builds the command line used to launch a child Karaf JVM by string concatenation, then executes it through\u00a0/bin/sh\u00a0(Unix) or\u00a0cscript\u00a0(Windows). The caller-supplied\u00a0javaOpts\u00a0value is spliced into that string unquoted. A javaOpts value containing shell metacharacters (;,\u00a0|,\u00a0`,\u00a0$(...)) is interpreted by the shell instead of being passed to the JVM as an option, giving arbitrary OS command execution as the Karaf process user.\n\n\nReachable via the shell commands\u00a0instance:create,\u00a0instance:start,\u00a0instance:restart,\u00a0instance:change-opts, and the equivalent\u00a0InstanceMBean\u00a0JMX operations (createInstance,\u00a0startInstance,\u00a0changeJavaOpts,\u00a0cloneInstance).\n\nMitigation\u00a0 * Set\u00a0karaf.secured.command.compulsory.roles=admin\u00a0in\u00a0etc/system.properties\u00a0to close the fail-open gap for all unconfigured command scopes.\n * Restrict which principals can reach\u00a0instance:*\u00a0commands and\u00a0InstancesMBean\u00a0via\u00a0etc/users.properties\u00a0role assignments.\n * Treat\u00a0javaOpts\u00a0passed to instance:create/instance:start/instance:change-opts/InstancesMBean\u00a0as untrusted input only from fully-trusted operators.",
"id": "GHSA-2wpj-35wr-vch7",
"modified": "2026-09-28T15:31:49Z",
"published": "2026-09-28T12:31:05Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-91006"
},
{
"type": "WEB",
"url": "https://lists.apache.org/thread/olzy0yjw82b20w59vonfjr1x7v5yzocr"
},
{
"type": "WEB",
"url": "http://www.openwall.com/lists/oss-security/2026/09/28/3"
}
],
"schema_version": "1.4.0",
"severity": []
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.