Description

In JetBrains Datalore before 2026.4 improper owner validation allowed reading another workspace's issues and runs notebook

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "4.3",
  "pubDate": "2026-11-06T11:17:07.730Z",
  "pubdate": "2026-11-06T11:17:07.730Z",
  "executiveSummary": "A critical vulnerability exists in JetBrains Datalore prior to version 2026.4 involving improper owner validation during resource access requests.\nThe vulnerability allows an authenticated attacker to bypass authorization controls and access sensitive information belonging to separate workspaces.\nAffected entities include notebook runs and associated issue tracking data stored within the Datalore environment.\nThis flaw represents a significant risk to data confidentiality, as it permits unauthorized cross-workspace data exfiltration.\nThe vulnerability requires an authenticated attacker to have access to the platform; however, no elevated administrative privileges are required to exploit the lack of proper object-level authorization checks.\nThe primary impact involves the unauthorized exposure of proprietary data, internal research artifacts, and potentially sensitive configuration information contained within notebook runs.",
  "technicalDetails": "The root cause of the vulnerability lies in an insufficient implementation of object-level access control (OLAC) within the Datalore backend service responsible for managing workspace resources.\nSpecifically, the application logic responsible for retrieving notebook run metadata and associated issue tracking data fails to verify that the requesting user possesses the necessary ownership or collaborator permissions for the target workspace resource.\nDuring the processing of API requests for workspace-specific data, the system relies on user-supplied identifiers (such as workspace IDs or resource GUIDs) without cross-referencing these identifiers against the session-authenticated user's authorization matrix.\nThe attack flow initiates when an authenticated attacker identifies the resource ID of a notebook run or issue belonging to a target workspace to which they have no legitimate access.\nBy crafting a request targeting these specific resources, the attacker can submit the request to the Datalore API endpoints that lack proper owner validation.\nBecause the server-side logic processes the request based solely on the presence of a valid session token, rather than verifying the binding between the authenticated user and the requested resource, the application returns the sensitive contents to the attacker.\nThis behavior affects various components of the Datalore platform, specifically impacting the reporting mechanisms for notebook execution logs and issue tracking modules.\nSince the authorization check is missing at the service layer, the application essentially operates as a passive information disclosure mechanism, allowing an attacker to enumerate and exfiltrate data across the multi-tenant environment.\nThis vulnerability is classified as an Insecure Direct Object Reference (IDOR) scenario where the application fails to enforce authorization checks on sensitive backend functions.\nThe lack of restrictive validation allows for automated scraping or targeted inspection of other users' workspace activities, leading to a complete compromise of workspace isolation guarantees.\nThe vulnerability is present in all JetBrains Datalore versions released prior to 2026.4, affecting both cloud-hosted and potentially on-premises instances where this specific resource management logic is deployed."
}