Database/AI/ML frameworks & serving
NVIDIA NeMo Export-Deploy: server-side request forgery in multimodal deployment
Impact
A server-side request forgery in the multimodal deployment path of NeMo Export-Deploy can make the model-serving process send requests the attacker chooses. NVIDIA rates the result as information disclosure. On a GPU node this matters because the serving process can usually reach addresses an outside client cannot, such as internal services, cluster-internal endpoints or cloud metadata. The record does not say which targets can be reached or what data comes back. The CVSS vector rates confidentiality impact as low.
Who can reach it
Anyone who can reach an exposed NeMo Export-Deploy multimodal endpoint over the network. No authentication or user interaction is needed (CVSS AV:N/PR:N/UI:N).
What to do
Update NeMo Export-Deploy to the release named in NVIDIA bulletin 2026/5920 and restart the serving deployment. The record does not give the fixed version. Until you patch, keep the endpoint off untrusted networks and limit the serving pod's outbound traffic to internal and metadata addresses.
References
Related entries
- vLLM: unauthenticated /tokenize caller sets video max_frames and fps, exhausting frontend memoryCVE-2026-105758 · vLLM Qwen2VLVideoBackend / Qwen3VLVideoBackend (request-level media_io_kwargs video limits)Medium
- vLLM: GLMGA video backend builds an attacker-sized frame-index list, starving the shared media loaderCVE-2026-105760 · vLLM GLMGA video backend (pre-decode frame-index list from request media_io_kwargs)Medium
- Open WebUI: missing authorization lets any logged-in user reach disabled model backends by indexCVE-2026-109296 · Open WebUI (Ollama and OpenAI-compatible model listing routes)Medium
- Crawl4AI: JavaScript injection via css_selector and wait_for bypasses the js_code blockCVE-2026-119121 · Crawl4AI Docker API server (css_selector / wait_for crawl parameters)Medium
- vLLM: model config.json can redirect the served model and tokenizer to another local checkpointCVE-2026-130684 · vLLM (speculators_config model path resolution)Medium
- BentoML OpenLLM 0.6.30 (async_run_command in src/openllm/common.py): A model repository directory name flows unescapedCVE-2026-15035 · BentoML OpenLLM 0.6.30 (async_run_command in src/openllm/common.py)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.