Automated Threat Modeling

Brief Summary: JetBrains TeamCity CVE-2026-132915 LDAP Injection Allowing Login Restriction Bypass

A brief summary of CVE-2026-132915, an LDAP injection (CWE 90) in JetBrains TeamCity that allowed bypassing login restrictions. It covers the likely root cause, the fixed builds 2026.2, 2026.1.4 and 2025.11.8, and how to confirm you are patched.

CVE Analysis

6 min read

Gaurav Sarraf
Gaurav Sarraf

2026-11-19

Brief Summary: JetBrains TeamCity CVE-2026-132915 LDAP Injection Allowing Login Restriction Bypass

Introduction

An LDAP injection flaw in JetBrains TeamCity let crafted login input get past the restrictions administrators place on who may sign in through LDAP. JetBrains shipped the fix in early September 2026, but the CVE only became public on 19 November 2026. Many teams are therefore learning about a High-severity authentication issue roughly 11 weeks after the patched builds were already available.

TeamCity is the continuous integration server from JetBrains, a company that has built developer tools since 2000. JetBrains says TeamCity is trusted by more than 30,000 DevOps teams, a vendor-supplied figure rather than an independent market measurement. Because TeamCity sits inside software build pipelines, it has drawn serious attention before. CISA documented nation-state exploitation of a different TeamCity flaw in 2023.

Technical information

What the vendor says

JetBrains' Fixed security issues bulletin describes the bug in one line: "An LDAP injection allowed bypassing login restrictions". The entry credits Than Kien as the reporter, lists the internal tracker ID TW-103102, rates the issue High, and classifies it as CWE-90 (Improper Neutralization of Special Elements used in an LDAP Query). The CNA record published through NVD scores it as follows:

CVSS 3.1: 7.1 AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:L/A:N

Read plainly, the vector says the issue is reachable over the network, has low attack complexity, requires low privileges, and needs no user interaction. It carries high confidentiality impact, low integrity impact and no availability impact.

Where user input meets LDAP

TeamCity is closed source and no commit, diff or code snippet is public, so we cannot show the vulnerable code. TeamCity's own documentation does, however, show where a typed username flows into LDAP during login:

  • teamcity.auth.loginFilter is a regular expression the typed username must match before any LDAP call is made. By default it rejects /, \ and @, which is why DOMAIN\username style logins are refused. When a username fails this check, the server log records that the entered login does not match teamcity.auth.loginFilter and that the attempt was aborted.
  • teamcity.users.login.filter is used when users live in different LDAP branches. On each login, TeamCity searches LDAP for the user with this filter using a service account (java.naming.security.principal / credentials), retrieves the user's DN, and then binds as that DN.

In practice, the username a person types ends up inside an LDAP search filter.

Likely root cause (our inference)

LDAP search filters give special meaning to characters such as *, (, ), \ and NUL. RFC 4515 requires these to be escaped when a value is meant to be treated literally. If a user-supplied login is inserted into a filter without that escaping, the input can alter which directory entries the filter matches. MITRE's description of CWE 90 covers the general pattern: malicious input changes the query and causes behavior the application did not intend.

The vendor's wording, "bypassing login restrictions", fits this picture. Administrators use these filters to constrain which users can authenticate, and a crafted login could get past those constraints. The standard remediation for CWE 90 is to escape every user-supplied value as a literal before it is placed in a filter. Stricter validation of the login before the search runs is also likely part of the change. This is our inference. JetBrains has not described its implementation, and the details live in the nonpublic ticket TW-103102.

Attack flow, as far as public data supports it

Based on the bulletin, the CVSS vector and the documented login path, the flow most consistent with the available information is:

  1. The attacker reaches the TeamCity login over the network (AV:N) and holds at least low privileges (PR:L), per the CNA vector.
  2. The attacker submits a login containing LDAP filter metacharacters.
  3. On an unpatched server using LDAP authentication, that value is placed into the search filter TeamCity runs with its service account.
  4. Because the value is not neutralized, the filter's logic changes, and the login restrictions the administrator configured no longer apply as intended.

Public material does not identify a specific request, payload, or the exact configuration required (for example, whether teamcity.users.login.filter must be in use). We include no exploit details here. The documented impact is limited to login restriction bypass, and nothing in the sources establishes code execution or server takeover.

Patch information

JetBrains fixed CVE-2026-132915 in its own TeamCity releases. There is no public source patch. Version numbers confirm the fix in three places: the vendor security bulletin, the CNA record on NVD, and JetBrains' download and release note pages. All three fixed builds shipped on the same day as one coordinated release:

  • 2026.2 (major release): released 2 September 2026
  • 2026.1.4, build 222865: released 2 September 2026
  • 2025.11.8, build 208382: released 2 September 2026

The 2026.1.4 release notes describe the security content only as "94 security problems have been fixed" and point readers to the bulletin. JetBrains says it usually publishes bulletins a few days after a release. In this case the CVE did not become public until 19 November 2026, about 11 weeks after the fixed builds shipped.

If you upgraded to any build released on or after 2 September 2026 on these lines, you already have the fix. Newer builds also include it, such as 2026.2.2 (build 239308) and 2026.1.6 (build 223074), both released 3 November 2026. JetBrains recommends installing those updates because they include security fixes.

Confirming you have the fix: the only reliable check is the server's version or build number. Compare it against the builds above. Later builds on each line also contain the fix.

Additional hardening (not a substitute for the upgrade):

  • OWASP recommends limiting the privileges assigned to the LDAP binding account. In TeamCity, that is the service account configured via java.naming.security.principal.
  • JetBrains' TeamCity security notes recommend complex passwords for users with administrative permissions, proper access policies, not enabling Guest Login, and deploying TeamCity in a trusted environment that malicious users cannot reach.

Affected systems and versions

According to the vendor bulletin and the CNA ranges published on NVD:

  • TeamCity 2026.1.x: versions from 2026.1 up to but not including 2026.1.4 are affected (CNA range 2026.1 < 2026.1.4). Fixed in 2026.1.4 (build 222865).
  • TeamCity 2025.11.x and earlier: all versions before 2025.11.8 are affected (CNA range 0 < 2025.11.8). Fixed in 2025.11.8 (build 208382).
  • TeamCity 2026.2: the major release contains the fix.

The vulnerable code path concerns LDAP-based login, so the documented exposure relates to servers that authenticate users against LDAP. The sources do not state precisely which LDAP configurations are required for exploitation.

Vendor security history

TeamCity has been exploited in the wild before through a different vulnerability. CISA and partner agencies reported that Russian Foreign Intelligence Service (SVR) actors, also known as APT29, exploited CVE-2023-42793 against internet-connected TeamCity servers at scale beginning in September 2023. That history is a good reason to keep TeamCity patched and to limit its network exposure, but it should not be read as evidence about this CVE.

For CVE-2026-132915 specifically, VulDB lists KEV status as no and classifies activity as very low, and we found no public report of active exploitation or threat actor attribution. Given the gap between the silent fix and public disclosure, the absence of reports is not proof that exploitation has not occurred.

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