Se Umea

To: abuse@umu.se Subject: Unauthenticated AI inference endpoint, Umeå University (130.239.40.121)

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

2026-05-01

Re: Unauthenticated Ollama AI inference endpoint, Umeå University IP / Host: 130.239.40.121 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

Umeå University (Sweden) has a named GPU compute server (gpuhost02.cs.umu.se) running Ollama with a large reasoning model (qwen3.6:35b) publicly accessible without authentication. Part of the Computer Science department compute cluster.

Infrastructure

Models

Findings

F1, Unauthenticated GPU Research Server (HIGH): Named GPU host #2 in CS compute cluster. All models injectable via CVE-2025-63389.

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/SE/umea.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