Description

In JetBrains PyCharm before 2026.3 code execution was possible in Safe Mode via a crafted Jupyter notebook widget module

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "5.5",
  "pubDate": "2026-11-06T17:17:24.577Z",
  "pubdate": "2026-11-06T17:17:24.577Z",
  "executiveSummary": "A critical vulnerability exists in JetBrains PyCharm versions prior to 2026.3, allowing for arbitrary code execution within the Integrated Development Environment (IDE).\nThe vulnerability manifests when the IDE operates in 'Safe Mode', a state intended to restrict functionality to prevent malicious code execution from untrusted projects.\nBy leveraging a maliciously crafted Jupyter notebook widget module, an attacker can bypass these security restrictions to execute arbitrary code with the privileges of the user running the IDE.\nThis vulnerability poses a significant risk to developers, as simply opening or interacting with a compromised notebook file can trigger the exploitation process.\nThe attack requires no authentication once the user opens the malicious project, effectively subverting the trust boundaries established by Safe Mode.\nSuccessful exploitation allows for complete compromise of the local development environment, potentially leading to unauthorized data exfiltration, local system persistence, or lateral movement within the developer's network.",
  "technicalDetails": "The root cause of this vulnerability lies in an improper validation and sandboxing mechanism within the Jupyter notebook widget integration in PyCharm. Safe Mode in PyCharm is designed to disable plugins and restrict the execution of untrusted scripts or build processes when an external or untrusted project is opened.\nHowever, the vulnerability indicates that the widget handling logic for Jupyter notebooks fails to sufficiently sanitize or restrict the execution context of specific module types, even when the IDE is configured in Safe Mode.\nThe exploitation process initiates when a victim opens a crafted Jupyter notebook containing a malicious widget module. In typical notebook environments, widgets are used to provide interactive UI elements; however, the PyCharm implementation fails to enforce strict execution boundaries during the deserialization or instantiation of these widgets.\nWhen the notebook is rendered or the widget is initialized by the IDE's internal Jupyter service, the malicious module triggers a code execution flow. Because the underlying mechanism does not effectively isolate the widget's execution context, the malicious code is executed directly within the context of the PyCharm application process.\nThe exploit flow follows these steps: 1) The attacker constructs a malicious `.ipynb` file containing a widget definition designed to exploit the parsing logic. 2) The victim opens this notebook within PyCharm. 3) PyCharm, while in Safe Mode, attempts to initialize the notebook UI components. 4) The Jupyter integration component processes the malicious widget, leading to an out-of-bounds or deserialization-based execution path. 5) Arbitrary commands are executed with the user's local shell or IDE-level permissions.\nAffected components include the internal Jupyter notebook support and the widget rendering engine within PyCharm versions prior to 2026.3. The vulnerability does not require authentication or user interaction beyond opening the malicious file, circumventing the intended security posture of Safe Mode."
}