CVE-2026-100254 in TeamCity
Summary
by MITRE • 09/30/2026
In JetBrains TeamCity before 2026.2, 2026.1.4, 2025.11.8 authenticated users could execute commands on Windows servers via CRLF injection in Pipeline Git connection settings
Be aware that VulDB is the high quality source for vulnerability data.
Analysis
by VulDB Data Team • 09/30/2026
The vulnerability identified in JetBrains TeamCity versions prior to 2026.2, specifically affecting the 2026.1.4 and 2025.11.8 releases, represents a critical security flaw rooted in improper input validation within the pipeline configuration subsystem. This issue allows authenticated users to execute arbitrary commands on Windows servers hosting TeamCity instances through CRLF injection attacks targeting Git connection settings. The core of this vulnerability lies in how the application processes user-supplied data for version control system configurations without adequately sanitizing special characters, specifically carriage return and line feed sequences. By injecting these control characters into fields intended for repository URLs or branch names, an attacker can manipulate the underlying command-line instructions that TeamCity uses to interact with Git repositories during build pipeline execution.
From a technical perspective, this flaw is classified under CWE-78 Improper Neutralization of Special Elements used in an OS Command, commonly known as OS Command Injection. The vulnerability exploits the fact that Windows batch scripts or PowerShell commands invoked by TeamCity may interpret newline characters as command terminators or separators. When an authenticated user provides a maliciously crafted Git connection string containing CRLF sequences, these characters can break out of the intended parameter context and append additional system commands to the execution pipeline. This effectively allows the attacker to bypass the sandboxed environment typically associated with build agents and execute arbitrary operating system-level instructions with the privileges assigned to the TeamCity service account or agent process on Windows hosts.
The operational impact of this vulnerability is severe, as it compromises the integrity and confidentiality of the continuous integration and deployment infrastructure. An authenticated user, which could range from a malicious insider to an external attacker who has compromised legitimate credentials through phishing or credential stuffing attacks, can gain unauthorized remote code execution capabilities. This access enables the exfiltration of sensitive source code, configuration files, and build artifacts stored on the server. Furthermore, it allows for lateral movement within the corporate network if the TeamCity agent possesses broader permissions than intended, potentially leading to full compromise of critical development environments and production systems dependent on these builds.
This attack vector aligns with MITRE ATT&CK technique T1059 Command and Scripting Interpreter, specifically sub-techniques involving Windows command shell or PowerShell execution via injection mechanisms such as CRLF injection which falls under input validation failures often associated with CWE-20 Improper Input Validation. The exploitation does not require privilege escalation beyond the existing authenticated user level but relies on the trust relationship between TeamCity and its build agents, highlighting a failure in principle of least privilege enforcement within pipeline configurations.
Mitigation strategies must prioritize immediate patching to the latest stable version of JetBrains TeamCity where this input validation flaw has been addressed through rigorous sanitization of special characters in Git connection fields. In environments where upgrading is not immediately feasible, administrators should implement strict network segmentation to isolate build agents from sensitive data stores and restrict outbound connections. Additionally, enforcing multi-factor authentication for all user accounts reduces the risk of unauthorized access leading to exploitation. Input validation controls at the web application firewall level can also help detect and block requests containing anomalous CRLF sequences in configuration endpoints, providing a temporary layer of defense until permanent remediation is applied.