Automated Threat Modeling

Brief Summary: Oracle MySQL Client Programs Flaw CVE-2026-114952 Allows Unauthenticated Data Read and Modification

A brief summary of CVE-2026-114952, a CVSS 9.1 flaw in the Client programs component of Oracle MySQL Server and MySQL Cluster that Oracle says an unauthenticated network attacker can exploit. Covers affected versions and Oracle's November 2026 patch information.

CVE Analysis

6 min read

Gaurav Sarraf
Gaurav Sarraf

2026-11-17

Brief Summary: Oracle MySQL Client Programs Flaw CVE-2026-114952 Allows Unauthenticated Data Read and Modification

Introduction

Oracle's off-cycle November 2026 Critical Security Patch Update fixes a flaw in the MySQL client programs that, according to Oracle, lets an unauthenticated attacker with network access read and modify data reachable by MySQL Server and MySQL Cluster. CVE-2026-114952 has a CVSS 3.1 score of 9.1, and its affected ranges extend to the newest GA builds on every supported line. That makes it relevant to nearly every current MySQL deployment, including hosts that only run the client tools.

Below we summarize what Oracle and third-party records say, where the public information stops, and what that means for patch planning.

Technical information

What Oracle has disclosed

The vulnerability sits in the Client programs component of MySQL Server and MySQL Cluster. Oracle's description states that an unauthenticated attacker with network access via multiple protocols can exploit it, and that successful attacks can result in:

  • unauthorized access to critical data or complete access to all MySQL Server and MySQL Cluster accessible data
  • unauthorized creation, deletion or modification of critical data or all accessible data

The CVSS 3.1 vector is CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:N. That means network reachable, low attack complexity, no privileges, no user interaction, unchanged scope, high confidentiality and integrity impact, and no availability impact.

Root cause: what is and is not known

Oracle has not published a commit, diff, or root cause note, and its CSAF record lists no CWE. The public material does not identify the protocols involved, the configuration prerequisites, or the exact data objects or operations exposed in a given deployment. Treat those details as unknown.

What we can say is that Client programs refers to the command-line tools that ship with the server, such as mysql, mysqldump, mysqlbinlog and mysqladmin. A client-side component combined with unauthenticated remote reachability usually means a client tool is trusting data sent by a server, or by an attacker positioned in the network path. If that holds, the fix most likely changes how those tools parse or validate what they receive. This is a hypothesis drawn from the component name and vector, and Oracle has not confirmed it.

Attack flow, as far as the public record supports

No proof of concept or exploitation writeup is public, so the following is limited to what the description and vector state:

  1. The attacker needs only network access to a reachable protocol endpoint. Oracle's main risk matrix names MySQL Protocol, while other Oracle records say multiple protocols.
  2. No credentials are required (PR:N) and no user interaction is required (UI:N), and Oracle rates the attack as easily exploitable (AC:L).
  3. A successful attack yields high confidentiality and integrity impact against data accessible to MySQL Server or MySQL Cluster, without affecting availability.

None of the sources we reviewed describe the specific trigger, the payload, or any timing or positioning requirements.

An inconsistency in Oracle's records

Oracle's main advisory risk matrix lists the attack protocol as MySQL Protocol, while the text-form risk matrix and the CSAF description say multiple protocols. The score, vector and version ranges are identical across all three sources, so this does not change what to patch. It does suggest the vulnerable code may be reachable through more than the classic MySQL wire protocol, which matters if you are planning network-level interim controls.

Exploitation status

The third-party vulnerability records we reviewed report that exploitation was not known at the time of writing, and they list no known exploit. We found no public proof of concept or named threat actor. Oracle credits Kentaro Takeda of YouMind with reporting the issue. That is reporter attribution, not evidence about any attacker. Treat this as point-in-time status rather than proof that exploitation has not occurred privately.

Patch information

Distribution. Oracle fixed CVE-2026-114952 in its November 2026 Critical Security Patch Update (CSPU), released on 17 November 2026. This is an off-cycle bundle on top of Oracle's usual quarterly Critical Patch Updates, and it contains 10 new security patches for Oracle MySQL. Oracle's machine-readable CSAF feed marks the remediation as vendor_fix and points to My Oracle Support Patch Availability Document CPU440, which is available to Oracle customers with valid support contracts. The fix ships as binary builds through Oracle's support channel, not as a public source diff.

Signs of the fixed builds. Oracle has not named fixed version numbers in the advisory. The public MySQL release notes now list new point releases on each affected line, all marked Not yet released as of this writing:

  • 8.4.13.1
  • 9.7.4.1
  • 26.10.0.1

These four-part version numbers match Oracle's usual pattern for off-cycle security rebuilds, and they are probably the builds that will carry this fix. That is an inference. No published release note yet ties CVE-2026-114952 to them. Until changelogs are published, CPU440 is the authoritative source for fixed versions.

Practical points.

  • Patch everywhere the client binaries run, not only on database servers. That includes application hosts, backup and replication tooling, CI runners and admin workstations.
  • Do not wait for a public source commit. Oracle typically pushes a squashed source drop to its public MySQL repository only when the community release ships, so there is no commit-level diff to review yet.
  • Interim controls. Oracle says blocking the network protocols required for an attack may reduce risk before patching, and a third-party record recommends limiting network access to trusted computers. Because the protocols are not fully identified, do not assume which traffic to block. Oracle warns that network blocking may break application functionality, should be tested on nonproduction systems, and does not fix the underlying issue or replace patching.

Public database status. At the time of writing, NVD returned CVE ID Not Found for this identifier, apparently because it has not been imported yet. Oracle's advisory, text-form risk matrix, and CSAF feed are the authoritative records.

Affected systems and versions

MySQL Server (Client programs component):

  • 8.4.0 through 8.4.13
  • 9.7.0 through 9.7.4
  • 26.10

MySQL Cluster:

  • 8.0.0 through 8.0.50
  • 8.4.0 through 8.4.13
  • 9.7.0 through 9.7.4
  • 26.10

Two details stand out:

  • MySQL Cluster 8.0.x is in scope, but MySQL Server 8.0 is not listed. The Server 8.0 line reached end of life with 8.0.46 in April 2026, so the fix on the 8.0 code line applies only to the Cluster product.
  • The newest builds are vulnerable. The top of each range (8.4.13, 9.7.4 and 26.10.0, all released 20 October 2026) is the latest GA build on that line. The fix is in releases after those.

References

Detect & fix
what others miss

Works with
  • GitHub
  • GitLab
  • Bitbucket
  • Azure DevOps Services
  • Gerrit
  • CVS
  • Jira
  • Linear
  • Slack
  • Security Compass
Explore integrations
Security magnifying glass visualization