Database/AI/ML frameworks & serving
NVIDIA OpenShell: CLI diagnostic command escaped for one shell layer but run through two, allowing command injection
Impact
On the remote path, OpenShell's command-line diagnostic interface passes user- or configuration-supplied values through two shell layers but escapes them for only one. Crafted values therefore run as commands, which NVIDIA says can lead to code execution, privilege escalation, information disclosure, data tampering and denial of service (scope changed). Attack complexity is high and user interaction is needed, so the likely path is getting an operator to run diagnostics against values the attacker controls.
Who can reach it
Local attacker with low privileges who can influence values or configuration that reach the diagnostic command. Requires a user to run the diagnostic (UI:R).
What to do
Update NVIDIA OpenShell CLI to the fixed release in NVIDIA product-security bulletin 5929. The record does not give a fixed version. Until then, avoid running the remote diagnostic command against sandboxes or configuration you do not trust.
References
Related entries
- MLflow: schemeless prompt source lets a caller read arbitrary files readable by the tracking serverCVE-2026-116690 · MLflow model registry (_create_model_version prompt branch / get-artifact)High
- vLLM orchestrator gateway: bearer tokens and full chat payloads written to persistent logsCVE-2026-15574 · vllm-orchestrator-gateway (Red Hat OpenShift AI)High
- SGLang (`/server_info`): Endpoint returns API keys and SSL keyfile pathsCVE-2026-15977 · SGLang (`/server_info`)High
- SGLang (weight exfiltration): Two endpoints allow a remote attacker to pull model weights when no API key is setCVE-2026-15978 · SGLang (weight exfiltration)High
- NVIDIA Triton Inference Server: unauthenticated request triggers excessive iteration and hangs the serverCVE-2026-16497 · NVIDIA Triton Inference Server for LinuxHigh
- Keras (HDF5 external links): Arbitrary local file read during model loadCVE-2026-1669 · Keras (HDF5 external links)High
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.