Am Armenian Academy
To: ipia@ipia.sci.am Subject: Unauthenticated AI inference endpoint, Institute for Informatics and Automation Problems, Armenia (37.26.168.19)
Nicholas Michael Kloster / NuClide Research nicholas@nuclide-research.com
2026-05-01
Re: Unauthenticated Ollama AI inference endpoint, Institute for Informatics and Automation Problems, Armenia IP / Host: 37.26.168.19 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
The Institute for Informatics and Automation Problems of the National Academy of Sciences of Armenia (Yerevan) is running Ollama inside a Docker container with two cloud proxy subscriptions. The 401 response leaks Docker container credentials, consistent with the Docker port-binding misconfig pattern also found at Hanoi University.
Infrastructure
| Field | Value |
|---|---|
| IP | 37.26.168.19 |
| rDNS | h168.019.yerphi.am |
| Org | Institute for Informatics and Automation Problems, NAS Armenia |
| Country | Armenia |
| Open ports | 11434 (Ollama, public) |
Credential Leak
{
"error": "unauthorized",
"signin_url": "https://ollama.com/connect?name=c2a68a9aa573&key=<base64>"
}
- Username:
c2a68a9aa573, Docker container ID - SSH pubkey:
ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIBqWiNKYbTt7XQxVG0OdY/61UHxsXkuGVtuS0UShBD7V
Both cloud proxy models return 401 with the same credentials, single Ollama Connect account inside one Docker container.
Models
| Model | Size | Type |
|---|---|---|
| deepseek-v4-pro:cloud | 0 GB | ☁️ Cloud proxy |
| minimax-m2.7:cloud | 0 GB | ☁️ Cloud proxy |
| llama3.2:3b | 1 GB | Local |
Findings
F1, Docker Credential Leak (HIGH): Container ID in Ollama Connect username.
F2, Dual Cloud Proxy on Academic Research Server (HIGH): DeepSeek and MiniMax subscriptions exposed.
F3, Model Injection (CRITICAL): All 3 models injectable via CVE-2025-63389.
Pattern
Docker container hostname as Ollama username also seen at: Hanoi University (Vietnam, 04aa6fb5e0b8), Purdue NW (US-IN, c0ddfaef7764). All three expose port 11434 via docker run -p 11434:11434 which binds to 0.0.0.0.
Why it matters
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/AM/armenian-academy.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