Description

In JetBrains Toolbox App before 3.6.1 nTLM hash disclosure was possible via SVG tool icons from a Remote Development server

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "6.5",
  "pubDate": "2026-11-06T17:17:24.880Z",
  "pubdate": "2026-11-06T17:17:24.880Z",
  "executiveSummary": "A vulnerability exists in JetBrains Toolbox App versions prior to 3.6.1 that allows for the unauthorized disclosure of NTLM hashes.\nThe issue stems from the way the application processes SVG tool icons retrieved from a Remote Development server.\nThis vulnerability is classified as an information disclosure issue, specifically an NTLM credential relay or capture vector.\nAn unauthenticated, remote attacker can trigger this vulnerability by enticing a user to connect to or interact with a malicious or compromised Remote Development environment.\nSuccessful exploitation enables an attacker to capture the victim's NTLMv2 hash, which can subsequently be used for offline brute-force attacks, pass-the-hash attacks, or credential relay to gain unauthorized access to internal network resources.\nThe risk is critical for enterprise environments where single sign-on or internal services rely on NTLM-based authentication.\nMitigation requires updating the JetBrains Toolbox App to version 3.6.1 or later to ensure proper sanitization and restricted handling of remote resources.",
  "technicalDetails": "The vulnerability resides within the JetBrains Toolbox App's rendering engine responsible for displaying metadata, specifically SVG-formatted tool icons, retrieved from Remote Development servers.\nWhen the application fetches these icons, it fails to sufficiently validate or restrict the network requests generated during the rendering process.\nIf an attacker controls a Remote Development server or can manipulate the metadata served to the client, they can specify a UNC path or a malicious URI scheme as the source for an SVG element.\nWhen the JetBrains Toolbox App attempts to resolve these resources, the underlying OS network stack may automatically attempt to authenticate the connection using the current user's security context.\nBecause the request originates from the local machine, the Windows operating system may initiate an NTLM authentication handshake with the attacker-controlled server to verify the user's identity.\nThe attack flow proceeds as follows: 1) The user connects to a malicious Remote Development server or views a compromised tool entry; 2) The application triggers a request for an external SVG icon; 3) The malicious resource points the application toward an attacker-controlled endpoint (e.g., via SMB or HTTP/WebDAV); 4) The local host transmits the NTLMv2 challenge-response pair to the attacker; 5) The attacker captures the hash for offline analysis or relaying.\nThis vulnerability is particularly dangerous because the NTLM exchange occurs transparently to the user, requiring no additional interaction beyond browsing or selecting the tool.\nThe scope of impact is restricted to environments where NTLM authentication is enabled and the application context allows network outbound traffic to arbitrary destinations.\nNo specific privileges are required on the client side, as the application processes the icons in the user's context automatically.\nThe lack of strict origin validation for resource loading in the icon rendering logic allows for potential SSRF-like behavior, where the application acts as a proxy to disclose authentication material to an untrusted third party."
}