Database/AI/ML frameworks & serving
JupyterLab: authenticated users bypass administrator plugin lock rules via /lab/api/plugins
Impact
Operators harden shared notebook environments by locking plugins off - stripping the extensions that would let a tenant move data off the node or widen the filesystem browser. Two server-side enforcement gaps let any user who already holds a notebook session re-enable or disable locked plugins by calling /lab/api/plugins directly, including child plugins of multi-plugin extensions and everything covered by the 'lock all' mechanism. The hardening an operator thought was enforced server-side is only enforced in the UI. This does not by itself grant code execution beyond what the notebook session already offers, but it removes controls such as upload and download limits that were the reason for locking.
Who can reach it
Any authenticated JupyterLab user with an existing notebook session on the node. No administrator rights are needed; the requests go straight to the server API and skip the UI that would have refused them.
What to do
Upgrade the jupyterlab package to 4.6.2 (4.6.x line) or 4.5.10 (4.5.x line) and restart each notebook server. Where notebooks run from a shared singleuser image, rebuild the image and have users restart their servers to pick it up. No node drain or reboot is involved.
References
Related entries
- Ray (dashboard DELETE endpoints): Browser-origin protection covers POST/PUT but not DELETECVE-2026-27482 · Ray (dashboard DELETE endpoints)Medium
- LocalAI (`/models/apply`): SSRF and partial local file inclusionCVE-2024-6095 · LocalAI (`/models/apply`)Medium
- NVIDIA NemoClaw: insufficiently protected credentials allow information disclosure and data tamperingCVE-2026-65087 · NVIDIA NemoClaw (credential storage)Medium
- Linux perf/x86/amd/uncore - memory leak in the events array: Per-CPU northbridge and last-level-cache uncore contextsCVE-2022-49784 · Linux perf/x86/amd/uncore - memory leak in the events arrayMedium
- PyTorch (flatbuffer loader): Out-of-bounds read parsing flatbuffer modelCVE-2024-31584 · PyTorch (flatbuffer loader)Medium
- Keras (HDF5 ExternalLink, incomplete fix): Arbitrary HDF5 file readCVE-2026-12480 · Keras (HDF5 ExternalLink, incomplete fix)Medium
This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.