GHSA-P54P-WJ4H-HHP9
Vulnerability from github – Published: 2026-09-15 09:30 – Updated: 2026-09-15 18:32A privilege escalation flaw was found in the scan_mysql actor of leapp-upgrade-el9toel10 (provided by leapp-repository). During RHEL 9 to RHEL 10 upgrades, the actor runs: mysqld --validate-config --log-error-verbosity=2 directly as root in the Leapp actor context, bypassing the packaged MySQL systemd unit that normally starts the daemon as User=mysql.
A process compromised as the mysql OS identity can write a version-2 persisted configuration (mysqld-auto.cnf) and a malicious shared object into /var/lib/mysql (a directory owned by mysql). That persisted map can set plugin_dir to /var/lib/mysql and early_plugin_load (or related loader options such as plugin_load / plugin_load_add) so MySQL loads the attacker-controlled object during configuration validation. Plugin loading can reach dlopen() before MySQL’s runtime-user check and before plugin-symbol validation.
When an administrator subsequently runs the documented Leapp preupgrade or upgrade workflow, attacker-controlled code can execute as UID 0 with a full capability set in an unconfined SELinux domain (unconfined_t). The attack does not require write access to the default system plugin path under /usr; redirecting plugin_dir via mysql-owned persisted state is sufficient. Ordinary SQL privileges alone (including highly privileged SQL accounts) are not a sufficient startpoint — OS-level execution as the mysql service identity is required, plus later administrator invocation of Leapp.
{
"affected": [],
"aliases": [
"CVE-2026-75092"
],
"database_specific": {
"cwe_ids": [
"CWE-250"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-09-15T09:16:43Z",
"severity": "HIGH"
},
"details": "A privilege escalation flaw was found in the scan_mysql actor of leapp-upgrade-el9toel10 (provided by leapp-repository). During RHEL 9 to RHEL 10 upgrades, the actor runs:\nmysqld --validate-config --log-error-verbosity=2 directly as root in the Leapp actor context, bypassing the packaged MySQL systemd unit that normally starts the daemon as User=mysql.\n\nA process compromised as the mysql OS identity can write a version-2 persisted configuration (mysqld-auto.cnf) and a malicious shared object into /var/lib/mysql (a directory owned by mysql). That persisted map can set plugin_dir to /var/lib/mysql and early_plugin_load (or related loader options such as plugin_load / plugin_load_add) so MySQL loads the attacker-controlled object during configuration validation. Plugin loading can reach dlopen() before MySQL\u2019s runtime-user check and before plugin-symbol validation.\n\nWhen an administrator subsequently runs the documented Leapp preupgrade or upgrade workflow, attacker-controlled code can execute as UID 0 with a full capability set in an unconfined SELinux domain (unconfined_t). The attack does not require write access to the default system plugin path under /usr; redirecting plugin_dir via mysql-owned persisted state is sufficient. Ordinary SQL privileges alone (including highly privileged SQL accounts) are not a sufficient startpoint \u2014 OS-level execution as the mysql service identity is required, plus later administrator invocation of Leapp.",
"id": "GHSA-p54p-wj4h-hhp9",
"modified": "2026-09-15T18:32:21Z",
"published": "2026-09-15T09:30:31Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-75092"
},
{
"type": "WEB",
"url": "https://github.com/oamg/leapp-repository/pull/1343"
},
{
"type": "WEB",
"url": "https://github.com/oamg/leapp-repository/commit/3e4cafffc8de1600a66519b326a407f37b665356"
},
{
"type": "WEB",
"url": "https://access.redhat.com/errata/RHSA-2026:67608"
},
{
"type": "WEB",
"url": "https://access.redhat.com/errata/RHSA-2026:67609"
},
{
"type": "WEB",
"url": "https://access.redhat.com/security/cve/CVE-2026-75092"
},
{
"type": "WEB",
"url": "https://bugzilla.redhat.com/show_bug.cgi?id=2517499"
},
{
"type": "WEB",
"url": "https://github.com/oamg/leapp-repository"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:R/S:U/C:H/I:H/A:H",
"type": "CVSS_V3"
}
]
}
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.