Ollama
Path Traversal leading to Remote Code Execution in Ollama
Ollama is vulnerable to path traversal in the `/api/pull` endpoint due to insufficient validation of layer digests by the `digestToPath` function. An unauthenticated remote attacker can specify a path traversal sequence as a layer digest, causing a malicious binary to be written outside the model store. Critically if the server process has write access to `/usr/lib/ollama` (the default in most Ollama Docker images), an attacker can write the malicious file to that directory. On the next server restart, the file is loaded and executed, resulting in remote code execution as root. This issue was fixed in version 0.35.0.
Affected versions and mitigation
From the CVE record and the vendor's references.Affected
- >= 0.14.3, < 0.35.0
Fixed in 0.35.0
Mitigation
Upgrade Ollama to 0.35.0 or later.
Timeline
Days relative to CVE publicationFirst public PoC 11 days after publication; not in CISA KEV. First PoC: striga-ai/CVE-2026-103663 (created 19 Oct 2026, found by us: 14 Nov 2026).
PoC repositories
Found on GitHub.1 PoCs
PoCs 1
| Repository | Created | Found by us | Stars | Assessment |
|---|---|---|---|---|
| striga-ai/CVE-2026-103663 | 19 Oct 2026 +11 d | 14 Nov 2026 | 0 | PoC |
Only PoCs count toward the first-PoC date and the characteristics below. Do not run PoC code from unknown sources.
PoC characteristics
What to hunt for in your logs · from 1 PoC repoGET /api/version
- Ports
- 11434
- Seen in
- 1 repo: striga-ai/CVE-2026-103663
- Evidence
- striga-ai/CVE-2026-103663/run.sh:21 striga-ai/CVE-2026-103663/run.sh:22 striga-ai/CVE-2026-103663/run.sh:74
POST /api/pull
- Body markers
"model":"{var}/x/y:latest""insecure""stream"intext/plain- Ports
- 11434
- Seen in
- 1 repo: striga-ai/CVE-2026-103663
- Evidence
- striga-ai/CVE-2026-103663/run.sh:63
Regex
Generated from PoC characteristics; not yet reviewed. Matches the request path and query.
(?i)^/api/version(?:[?#].*)?$
(?i)^/api/pull(?:[?#].*)?$
Suricata
Template generated from PoC characteristics. Review before deploying.
alert http any any -> $HOME_NET [11434] (msg:"GG PoC-Monitor CVE-2026-103663 Ollama Ollama exploit attempt"; flow:established,to_server; http.method; content:"GET"; http.uri; content:"/api/version"; reference:url,poc.griffinguard.io/cve/CVE-2026-103663; classtype:web-application-attack; sid:9273260; rev:1;)
alert http any any -> $HOME_NET [11434] (msg:"GG PoC-Monitor CVE-2026-103663 Ollama Ollama exploit attempt"; flow:established,to_server; http.method; content:"POST"; http.uri; content:"/api/pull"; http.request_body; content:"|22|model|22|:|22|"; http.request_body; content:"|22|insecure|22|"; http.request_body; content:"|22|stream|22|"; reference:url,poc.griffinguard.io/cve/CVE-2026-103663; classtype:web-application-attack; sid:9443789; rev:1;)
Sigma
Template generated from PoC characteristics. Review before deploying.
title: Ollama Ollama exploit attempt (CVE-2026-103663)
id: gg-pocmon-cve-2026-103663-9273260
status: experimental
description: Request pattern taken from public PoCs for CVE-2026-103663. Generated, not reviewed.
references:
- https://poc.griffinguard.io/cve/CVE-2026-103663
logsource:
category: webserver
detection:
selection:
cs-method: GET
cs-uri-stem|contains|all:
- '/api/version'
condition: selection
falsepositives:
- Authorized vulnerability scanning
level: critical
tags:
- attack.initial_access
- attack.t1190
- cve.2026.103663
title: Ollama Ollama exploit attempt (CVE-2026-103663)
id: gg-pocmon-cve-2026-103663-9443789
status: experimental
description: Request pattern taken from public PoCs for CVE-2026-103663. Generated, not reviewed.
references:
- https://poc.griffinguard.io/cve/CVE-2026-103663
logsource:
category: webserver
detection:
selection:
cs-method: POST
cs-uri-stem|contains|all:
- '/api/pull'
condition: selection
falsepositives:
- Authorized vulnerability scanning
level: critical
tags:
- attack.initial_access
- attack.t1190
- cve.2026.103663