Automated Threat Modeling

Quick Look: CVE-2026-49372 — Unauthenticated SSRF in JetBrains TeamCity Build Status

CVE-2026-49372 is an unauthenticated Server-Side Request Forgery vulnerability in the build status feature of JetBrains TeamCity On-Premises, with a CVSS score of 7.5. Any attacker who can reach the server can use it to send requests into the network behind it, which puts CI/CD infrastructure at risk.

CVE Analysis

10 min read

ZeroPath CVE Analysis
ZeroPath CVE Analysis

2026-05-29

Quick Look: CVE-2026-49372 — Unauthenticated SSRF in JetBrains TeamCity Build Status
Experimental AI-Generated Content

This CVE analysis is an experimental publication that is completely AI-generated. The content may contain errors or inaccuracies and is subject to change as more information becomes available. We are continuously refining our process.

If you have feedback, questions, or notice any errors, please reach out to us.

[email protected]

Introduction

An unauthenticated Server-Side Request Forgery vulnerability in the build status feature of JetBrains TeamCity lets any network-reachable attacker turn a CI/CD server into a proxy for internal network reconnaissance and cloud credential theft. TeamCity servers routinely hold source code, deployment credentials, and production access, and APT29 and ransomware operators have already targeted the product. CVE-2026-49372 adds a meaningful new entry point to a threat surface that attackers are already working.

Technical Information

Root Cause

CVE-2026-49372 is classified under CWE-918 (Server-Side Request Forgery). The vulnerability is in TeamCity's build status feature, which reports and queries the status of build executions. In CI/CD workflows, build status endpoints are commonly designed to be reachable from external systems, so that chat platforms, dashboards, and deployment pipelines can integrate with them.

The build status endpoint accepts URL input (likely for build status callbacks or notifications) and uses it to make server-side HTTP requests without properly validating the user-supplied URL. In the terms of MITRE's CWE-918 definition, the application fetches a remote resource without making sure the destination is safe or expected. The endpoint also requires no authentication of any kind, so any attacker who can reach the TeamCity server over the network can trigger the SSRF.

Attack Flow

Going by the CWE-918 framework and the documented characteristics of this vulnerability, exploitation would follow this general sequence:

  1. Reconnaissance: The attacker finds a TeamCity On-Premises instance reachable over the network. Many TeamCity deployments are internet-facing to support remote development and CI/CD operations.

  2. Crafting the SSRF request: The attacker sends a request to the build status endpoint with a URL pointing to an internal resource. No authentication credentials are needed.

  3. Server-side request execution: The TeamCity server processes the request and opens an HTTP connection to the URL the attacker chose. Because the request comes from the TeamCity server itself, it gets past the network segmentation controls and firewalls that would normally block external access to internal services.

  4. Exploitation scenarios:

    • Internal network scanning: The attacker targets internal IP addresses and ports to find live hosts and running services. Response timing and status codes reveal network topology even when the SSRF is blind.
    • Cloud metadata access: In cloud-hosted deployments (AWS, GCP, Azure), the attacker targets the metadata endpoint at 169.254.169.254 to retrieve instance credentials, IAM tokens, and configuration data. Those credentials can enable privilege escalation within the cloud environment.
    • Credential harvesting: If the SSRF returns response content to the attacker (non-blind SSRF), the attacker can exfiltrate sensitive data from internal service responses directly.
    • Chaining with internal vulnerabilities: The SSRF becomes a pivot to internal services such as databases, message queues, or internal APIs, which may have exploitable weaknesses of their own.

CVSS and Severity Context

The assigned CVSS score of 7.5 is in the High severity range, which fits an unauthenticated, network-exploitable vulnerability that affects confidentiality. The NVD record is currently marked as "Undergoing Enrichment," so the full CVSS vector string and CPE applicability statements have not yet been finalized. The final score may be adjusted once enrichment is complete.

Relationship to CVE-2026-44413

CVE-2026-49372 was published on the same day as CVE-2026-44413, a high-severity post-authentication vulnerability that lets authenticated users (including guest accounts, if guest access is enabled) expose parts of the TeamCity server API to unauthorized access. The same versions (2026.1 and 2025.11.5) fix both vulnerabilities, and both affect only TeamCity On-Premises.

The difference between them is authentication. CVE-2026-49372 needs none at all, so attackers can exploit it at scale with automated scanning tools. CVE-2026-44413 needs at least a guest-level account, but once an attacker has any foothold it potentially opens up broader API access. Organizations should patch both at the same time.

AttributeCVE-2026-49372CVE-2026-44413
Vulnerability TypeSSRF (CWE-918)Auth bypass / API exposure (CWE-306)
Authentication RequiredNoYes (any authenticated user, including guest)
CVSS Score7.5High (undergoing NVD enrichment)
Attack VectorBuild status endpointServer API endpoints
Fixed In2026.1, 2025.11.52026.1, 2025.11.5
TeamCity Cloud AffectedNoNo

Architectural Pattern

This vulnerability fits a recurring pattern in TeamCity's architecture. As a CI/CD orchestrator, TeamCity has to make outbound network connections to repositories, build agents, notification systems, and deployment targets. That requirement leaves a persistent attack surface for SSRF, and five distinct SSRF-classified CVEs have been documented since 2020: CVE-2020-35667, CVE-2021-31910, CVE-2022-24333, CVE-2022-46830, and now CVE-2026-49372. Each one has been patched, but the recurrence suggests the root cause is architectural rather than incidental. Because the product legitimately needs outbound connectivity, every point that accepts a URL or network destination from user-controlled input has to validate it, and that remains an ongoing challenge.

Affected Systems and Versions

The following TeamCity On-Premises versions are affected:

  • All JetBrains TeamCity On-Premises versions before 2026.1
  • All JetBrains TeamCity On-Premises versions before 2025.11.5 (in the 2025.x release line)

The vulnerability is fixed in:

  • TeamCity 2026.1
  • TeamCity 2025.11.5

TeamCity Cloud is not affected. JetBrains has confirmed that TeamCity Cloud deployments need no action for this vulnerability. Organizations should check whether their deployment is On-Premises or Cloud before assessing their exposure.

The vulnerability affects every TeamCity installation where the firewall permits inbound connections to the TeamCity server, particularly internet-facing ones. Installations where build agents run on the same host as the TeamCity server may have additional exposure.

Vendor Security History

JetBrains TeamCity has a notable record of security vulnerabilities, several of which reached Critical severity and were exploited in the wild. The table below lists key CVEs from the JetBrains security fixes page:

CVEDescriptionSeverityFixed In
CVE-2026-44413Authenticated users could expose server API to unauthorized accessHigh2026.1, 2025.11.5
CVE-2026-28194Open redirect in React project creation flowMedium2025.11.3
CVE-2026-28195Missing authorization for project developers to add parametersMedium2025.11.3
CVE-2024-27198Authentication bypass allowing admin actionsCritical2023.11.4
CVE-2024-23917Authentication bypass leading to RCECritical2023.11.3
CVE-2023-42793Authentication bypass leading to RCE on TeamCity ServerCritical2023.05.4
CVE-2023-34218Bypass of permission checks allowing admin actionsCritical2023.05, 2022.10.4
CVE-2022-46830SSRF via custom STS endpointMedium2022.10.1
CVE-2022-25263OS command injection in Agent Push featureHigh2021.2.3
CVE-2022-24333Blind SSRF via XML-RPC callMedium2021.2
CVE-2021-31910Information disclosure via SSRFHigh2020.2.3
CVE-2020-35667Plugin SSRF potentially exposing user credentialsHigh2020.2.85695

Authentication bypass and SSRF vulnerabilities keep recurring across multiple years and major versions. Organizations running TeamCity On-Premises should patch on a rigorous schedule and keep defense-in-depth controls around their CI/CD infrastructure.

State-sponsored threat actors have targeted JetBrains TeamCity directly. APT29 (Cozy Bear) exploited CVE-2024-27198 in real-world campaigns. Trend Micro documented exploitation of TeamCity vulnerabilities that led to deployments of Jasmin ransomware, XMRig cryptocurrency miners, Cobalt Strike beacons, and SparkRAT backdoors. Darktrace observed rapid exploitation of CVE-2024-27198 across customer environments. According to Vectra AI, SSRF attacks overall surged 452% in 2024, driven in part by AI-powered automation of vulnerability discovery and exploitation.

As of May 29, 2026, no public proof-of-concept exploit or active exploitation of CVE-2026-49372 has been documented. Exploitation attempts still look likely in the near term: the vulnerability needs no authentication, CI/CD infrastructure is a high-value target, and sophisticated threat actors have already shown interest in TeamCity.

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