GHSA-pppj-hq3g-57pj: JupyterLab: crafted settings file executes code

JupyterLab 4.5+ allows notebook settings to be shared and applied through an `overrides.json` file using the `Import` button in the Settings Editor. Certain notebook display settings were not properly validated before being applied. As a result, a crafted settings file could contain hidden instructions that run as code inside JupyterLab when imported, instead of only changing a display preference. Because importing a settings file appears harmless, a user could import a file shared by another party without realizing it could do more. On multi-tenant file systems without proper permission control, another user could plant a malicious `overrides.json`. > CVE assignment pending, GitHub CNA is experiencing severe backlog ### Impact When a malicious settings file is applied, the embedded code runs with the same access as the affected user. This could allow an attacker to read or modify that user's notebooks and files, and to run code on the user's behalf through the notebook server, including on any connected kernel. #### User Interaction vs Privileges Required ##### Write access to a loaded settings location If an attacker can write to a directory JupyterLab loads settings from (e.g. on shared or multi-tenant file system), they could place a crafted `overrides.json` that is applied to another user automatically at startup. This requires high privilages but no action by the victim. ##### User-imported settings file A user can import a crafted `overrides.json` through the `Import` button in the Settings Editor, having received it from another party. This requires no privileges but a deliberate action by the victim, who reasonably expects a settings file to change preferences rather than run code. ### Patches JupyterLab 4.6.2 and 4.5.10 were patched. ### Workarounds None ### Hardening 1. Treat a settings file as something that can affect how JupyterLab behaves, not only how it appears. Administrators are encouraged to establish a trusted process for distributing configuration rather than relying on ad-hoc importing of shared files. 2. On multi-tenant or shared file systems, restrict write permissions on the application settings directory and other Jupyter configuration paths so that one user cannot place an `overrides.json` (or other configuration) readable by another user. A settings file in these locations is applied automatically, without an import step, so directory permissions are the primary control against cross-user tampering.

An attacker crafts a overrides.json disguised as a team 'dark theme' or 'standard settings pack' and shares it via email or a team chat channel. A data scientist, expecting only a cosmetic change, imports it through the Settings Editor's Import button; the embedded code executes with the user's privileges, giving the attacker the ability to read notebooks containing proprietary model code and API keys, modify files, or issue commands to any connected kernel - for example exfiltrating training data or planting a backdoor in a training script. In the multi-tenant variant, an attacker with a low-privileged account on a shared JupyterHub server writes the crafted overrides.json to a settings path loaded automatically for other tenants, achieving silent code execution against every user who next starts their environment, with no click required.