Ca Mb U Manitoba

To: infosec@umanitoba.ca Subject: Unauthenticated AI inference endpoint, University of Manitoba (130.179.30.15)

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

2026-05-01

Re: Unauthenticated Ollama AI inference endpoint, University of Manitoba IP / Host: 130.179.30.15 Severity: HIGH

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

The Computer Science department at the University of Manitoba (quail.cs.umanitoba.ca) is running Ollama with five large local models including DeepSeek-R1:70B, Llama 3.3, and Llama 3:70B, a deep research stack totaling ~156GB of local models, all accessible without authentication.

Infrastructure

Models

Total local compute: ~133 GB across 5 models.

Findings

F1, Unauthenticated CS Research Server (HIGH): Named GPU server in CS department. Research models (DeepSeek-R1, large Llama) and code model (Qwen2.5-Coder) exposed to the public internet.

F2, Model Injection (HIGH): All 5 models injectable via CVE-2025-63389, attacker can overwrite system prompts, affecting any research workflows using this Ollama instance.

Why it matters

Any internet actor can run uncapped inference against your GPU at your compute cost, and inject malicious system prompts into any loaded model via CVE-2025-63389.

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/CA/MB-u-manitoba.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