Automated Threat Modeling

Brief Summary: JetBrains TeamCity CVE 2026 132234, Mercurial VCS Root Argument Injection Leading to Remote Code Execution

A brief summary of CVE 2026 132234, a CVSS 9.9 argument injection flaw in JetBrains TeamCity Mercurial VCS root settings that enabled remote code execution. Covers the root cause, fixed builds, affected versions, and the vendor's TeamCity security history.

CVE Analysis

7 min read

Gaurav Sarraf
Gaurav Sarraf

2026-11-18

Brief Summary: JetBrains TeamCity CVE 2026 132234, Mercurial VCS Root Argument Injection Leading to Remote Code Execution

Introduction

A critical argument injection flaw in JetBrains TeamCity allowed a user holding only low privileges to turn Mercurial VCS root settings into remote code execution on the CI/CD server. TeamCity sits at the center of build and delivery workflows. An issue rated 9.9, with changed scope and high confidentiality, integrity, and availability impact, warrants a prompt review of every installation, even though the fix has been available since July 2026.

TeamCity is the continuous integration server from JetBrains, which the vendor describes as an advanced CI/CD solution supporting flexible workflows and collaboration. JetBrains, founded in 2000, builds tools for software developers and teams and reports that its products are used by individual developers, startups, educational institutions, and large organizations. Our sources do not establish TeamCity's market share, but CISA's past reporting on exploitation of TeamCity by state actors shows that attackers consider these servers valuable targets.

Technical information

Root cause

The CVE record states that in JetBrains TeamCity before 2026.1.3 and 2025.11.7, remote code execution was possible via argument injection in Mercurial VCS root settings. It classifies the weakness as CWE 88 (Improper Neutralization of Argument Delimiters in a Command, commonly called Argument Injection).

The mechanics follow from how TeamCity talks to Mercurial. For a Mercurial VCS root, the TeamCity server or agent runs the Mercurial client hg, and values taken from the VCS root settings become arguments to that command. CWE 88 means that some setting could carry text that hg reads as an extra option rather than as plain data. A value beginning with a dash character (single or double) is the classic example, and some Mercurial options can make it run arbitrary commands. That is the path from "edit a VCS root field" to "code running on the server".

JetBrains tracks the issue internally as TW 101698. The YouTrack ticket is not publicly viewable, so the specific vulnerable field, the Mercurial option involved, and the code change are not available for review.

CVSS profile

The record assigns a CVSS 3.1 base score of 9.9 (Critical) with the vector AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H. In practical terms:

  • AV:N, AC:L: reachable over the network with low attack complexity.
  • PR:L: the attacker needs an account with low privileges, so the bug requires authentication. The record does not specify which TeamCity role or permission satisfies the requirement.
  • UI:N: no action from another user is needed.
  • S:C: scope is changed, which fits a permission at the settings level turning into execution on the underlying server.
  • C:H, I:H, A:H: high impact to confidentiality, integrity, and availability.

Attack flow

Based on the published description and the weakness class, exploitation would proceed roughly as follows:

  1. The attacker obtains enough access to TeamCity, at a low privilege level, to edit Mercurial VCS root settings.
  2. The attacker places a value in a VCS root setting, crafted so hg will interpret it as a command option instead of as data.
  3. TeamCity (server or agent) builds and executes an hg command line that includes the attacker's value as an argument.
  4. hg processes the injected option. Because certain Mercurial options can execute arbitrary commands, the attacker gains code execution on the host running the command.

The precise argument, configuration field, and payload are not published, so we cannot reconstruct a concrete exploit sequence beyond this outline.

Exploitation status

The CVE record does not state that this issue has been exploited and names no threat actor. The CISA Known Exploited Vulnerabilities catalog text we reviewed did not list it, and we found no public proof of concept, indicators of compromise, or exploitation telemetry. Active exploitation is therefore unconfirmed. That absence is not proof that exploitation has not occurred, and prior attacks against other TeamCity flaws should not be read as evidence about this one.

Patch information

Where the fix ships. JetBrains has not released a source diff, commit, or standalone hotfix for CVE 2026 132234. The fix exists only inside two TeamCity On Premises bugfix builds, both released on 27 July 2026:

  • 2026.1.x line: fixed in 2026.1.3, build 222742. Affected range per the CNA: 2026.1 up to, but not including, 2026.1.3.
  • 2025.11.x line and earlier: fixed in 2025.11.7, build 208264. Affected range per the CNA: everything below 2025.11.7.

The CVE record was published on 18 November 2026, almost four months after these builds came out. That gap reflects how JetBrains normally works. It ships security fixes quietly in a bugfix release, and the release notes say only that the bulletin usually follows "a few days after the release date." In practice, anyone who updated to 2026.1.3 or 2025.11.7 over the summer already has this fix.

How JetBrains tracks it. Filtering the JetBrains Fixed Security Issues bulletin to TeamCity 2026.1.3 shows this issue under tracker ID TW 101698, with severity Critical, weakness CWE 88 (Argument Injection), and fixed versions 2026.1.3 and 2025.11.7. The fix is bundled with many others. The 2026.1.3 release notes report "30 security problems have been fixed," and the 2025.11.7 notes report 22. You cannot apply this fix in isolation. The whole update must be installed.

What the fix has to address. Fixes for this class of bug usually do one or more of the following:

  • place two consecutive dashes, which mark the end of options, before values the user supplies, so hg cannot read them as flags
  • reject or sanitize settings that start with a dash, or that contain characters Mercurial treats specially
  • check values such as the repository URL and paths against an allowlist of expected formats before building the command

These are the usual techniques for this weakness class. They are not confirmed details of TW 101698, and JetBrains has not said which approach it took.

Applying the update. The JetBrains release announcement notes that administrators can use the TeamCity automatic update feature, download the update from JetBrains, or pull the updated TeamCity Docker image. JetBrains strongly recommends upgrading as soon as possible. JetBrains has published no approved workaround specific to this CVE, so do not treat an unverified configuration change as a substitute for the update.

Checking that you are patched. The build number is the only thing you can verify to confirm the fix level. Look in the footer of the TeamCity web UI or under Administration, then Server Administration. Build 222742 or later on the 2026.1 line, or 208264 or later on the 2025.11 line, means the fix is present. Later releases on either line (for example 2026.1.4 or 2025.11.8) also include it, because fixes carry forward within a release line.

If you suspect compromise. JetBrains has separately published guidance on investigating a compromised TeamCity On Premises server. It recommends considering stopping unknown background processes and notes that the TeamCity Audit screen records configuration changes and security actions. This guidance predates the CVE and is not specific to it, but the Audit screen is a natural place to review recent VCS root edits.

Affected systems and versions

  • Product: JetBrains TeamCity (On Premises).
  • 2026.1 release line: versions 2026.1 up to, but not including, 2026.1.3. Fixed in 2026.1.3 (build 222742).
  • 2025.11 release line and earlier: all versions below 2025.11.7. Fixed in 2025.11.7 (build 208264).
  • Vulnerable component: Mercurial VCS root settings. The record does not state that other version control integrations are affected.
  • Preconditions: an attacker with low privileges (exact role unspecified) and network access to the TeamCity instance. The available sources do not establish whether a given deployment's configuration changes its exposure.

Vendor security history

TeamCity has a documented history of serious, distinct vulnerabilities, which should be kept separate from this Mercurial issue:

  • CVE 2023 42793: CISA reported that this flaw enabled authorization bypass and arbitrary code execution on TeamCity servers, and that APT29 (the Russian Foreign Intelligence Service, SVR) exploited it beginning in late September 2023.
  • CVE 2024 23917: in February 2024, CISA issued an alert for this separate TeamCity On Premises issue, warning that a threat actor could take control of an affected system and recommending updates or workarounds.
  • In March 2024, JetBrains published guidance on investigating compromised TeamCity On Premises servers and a post describing its approach to addressing the vulnerabilities disclosed at that time.

The 2026.1.3 and 2025.11.7 releases are more maintenance updates focused on security, and JetBrains states that each addressed more than 20 security vulnerabilities. Those counts do not by themselves describe the severity or exploitation status of each included issue. None of this history establishes exploitation or actor involvement for CVE 2026 132234.

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