Ollama versions 0.14.0 before 0.31.2 contain an incorrect authorization vulnerability in the experimental agent mode Bash tool approval mechanism… (CVE-2026-102697)
Ollama versions from 0.14.0 up to but not including 0.31.2 have an incorrect authorization vulnerability in the experimental agent mode Bash tool approval mechanism. This flaw allows attackers who can influence model output via prompt injection to append shell control operators to approved commands, enabling execution of additional shell commands without session approval. The vulnerability has a high severity with a CVSS score of 7.8.
AI Analysis
Technical Summary
The vulnerability in Ollama versions 0.14.0 before 0.31.2 involves improper parsing of shell syntax in the experimental agent mode Bash tool approval mechanism. This incorrect authorization allows an attacker who can manipulate model output through prompt injection to bypass session approval by appending shell control operators such as semicolons or logical operators to commands that are otherwise approved. This leads to unauthorized execution of arbitrary shell commands.
Potential Impact
Successful exploitation can result in the execution of arbitrary shell commands with the privileges of the affected process, potentially leading to full compromise of confidentiality, integrity, and availability of the system running the vulnerable Ollama version.
Mitigation Recommendations
A fix is available in Ollama version 0.31.2 and later. Users should upgrade to version 0.31.2 or newer to remediate this vulnerability. No other mitigation guidance is provided.
Ollama versions 0.14.0 before 0.31.2 contain an incorrect authorization vulnerability in the experimental agent mode Bash tool approval mechanism… (CVE-2026-102697)
Description
Ollama versions from 0.14.0 up to but not including 0.31.2 have an incorrect authorization vulnerability in the experimental agent mode Bash tool approval mechanism. This flaw allows attackers who can influence model output via prompt injection to append shell control operators to approved commands, enabling execution of additional shell commands without session approval. The vulnerability has a high severity with a CVSS score of 7.8.
CVSS v3.1
Score 7.8high
Affected software
pkg:github/ollama/ollamaRun on your own infrastructure? Check whether these packages are installed with threat-finder — our free open-source scanner.
Weaknesses
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
The vulnerability in Ollama versions 0.14.0 before 0.31.2 involves improper parsing of shell syntax in the experimental agent mode Bash tool approval mechanism. This incorrect authorization allows an attacker who can manipulate model output through prompt injection to bypass session approval by appending shell control operators such as semicolons or logical operators to commands that are otherwise approved. This leads to unauthorized execution of arbitrary shell commands.
Potential Impact
Successful exploitation can result in the execution of arbitrary shell commands with the privileges of the affected process, potentially leading to full compromise of confidentiality, integrity, and availability of the system running the vulnerable Ollama version.
Mitigation Recommendations
A fix is available in Ollama version 0.31.2 and later. Users should upgrade to version 0.31.2 or newer to remediate this vulnerability. No other mitigation guidance is provided.
Technical Details
- Gcve Source
- db.gcve.eu
- Osv Id
- GHSA-44m8-pr79-3734
- Osv Schema Version
- 1.4.0
- Aliases
- ["CVE-2026-102697"]
- Database Specific Severity
- HIGH
- Cvss Version
- 3.1
Threat ID: 6abc27c0680226ef6846f800
Added to database: 09/29/2026, 21:04:00 UTC
Last enriched: 09/29/2026, 21:16:58 UTC
Last updated: 09/30/2026, 03:12:02 UTC
Views: 6
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.