Vn Vnu Hcmc

To: info@vnuhcm.edu.vn Subject: Unauthenticated AI inference endpoint, Vietnam National University Ho Chi Minh City (103.88.123.165)

Nicholas Michael Kloster / NuClide Research nicholas@nuclide-research.com

2026-05-01

Re: Unauthenticated Ollama AI inference endpoint, Vietnam National University Ho Chi Minh City IP / Host: 103.88.123.165 Severity: CRITICAL

I’m an independent security researcher. I hold CISA disclosures CVE-2025-4364 and ICSA-25-140-11 and conduct good-faith AI infrastructure research under the NuClide Research umbrella. This is an unsolicited disclosure, no engagement exists with your organization, and I have not accessed, modified, or exfiltrated any data beyond what was necessary to confirm the exposure.

Summary

Vietnam National University Ho Chi Minh City (Information Technology Park) has an Ollama instance with an unusually named model final-exploit-v1:latest and a gpt-oss:latest cloud proxy. The final-exploit-v1 model is 168 bytes, the size of a cloud proxy artifact, not a local model, suggesting a custom-named cloud redirect.

Infrastructure

Models

Findings

F1, Unauthenticated Ollama API (HIGH): All models accessible without credentials.

F2, final-exploit-v1 Cloud Proxy Artifact (MEDIUM): The model is 168 bytes, identical size pattern to standard cloud proxy models (deepseek-v4-pro:cloud = 344 bytes, minimax = 384 bytes). Modified January 2026. No system prompt returned. Inference returns empty response. Likely a student-created cloud proxy with custom naming.

F3, gpt-oss Cloud Proxy (HIGH): gpt-oss:latest 12GB present. Status (200 OK vs 401) not confirmed on final probe.

F4, Model Injection (HIGH): All models injectable via CVE-2025-63389.

Why it matters

Any internet actor can run inference against your cloud API subscription at your expense, this constitutes direct quota/billing theft. The credential leak (username + SSH public key) exposes your service account to enumeration and credential-stuffing against other services.

One-line fix

OLLAMA_HOST=127.0.0.1:11434
systemctl restart ollama

This rebinds Ollama to loopback only. If running in Docker: docker run -p 127.0.0.1:11434:11434 ollama/ollama.

CVE-2025-63389

All models on this instance are injectable via the unauthenticated /api/create endpoint, an attacker can overwrite any model’s system prompt or delete models entirely. No patch exists as of this disclosure.

Reference

Full technical details, parameter counts, and remediation notes are in this public research repository: AI-LLM-Infrastructure-OSINT/blob/main/case-studies/universities/VN/vnu-hcmc.md

This research is part of a broader sweep of university AI infrastructure exposures documented at: AI-LLM-Infrastructure-OSINT/blob/main/case-studies/universities/OVERVIEW.md

I’m happy to answer questions or assist with verification. No response is required.

Regards, Nicholas Michael Kloster / NuClide Research nicholas@nuclide-research.com AI-LLM-Infrastructure-OSINT