CVE-2026-0313: Cortex OS Query content pack command injection

What it is

The OS Query content pack for Cortex XSOAR and Cortex XSIAM lets playbooks run queries on remote endpoints over SSH. A command injection flaw (CWE-78) in the pack lets an authenticated administrator run arbitrary OS commands on those remote endpoints.

This is a post-authentication bug. The attacker needs a high-privileged account in the platform first. Its value is reach: one admin account in XSOAR or XSIAM becomes command execution on every endpoint the SSH integration can connect to. Palo Alto Networks rates it medium (CVSS-BT 6.1) with a suggested urgency of moderate, though the base score is 8.6 (CVSS 4.0). Exploit maturity is listed as unreported.

Palo Alto Networks credits Mohamed Lemine Ahmed Jidou (AegisSec) for reporting it. The vendor says it is not aware of any malicious exploitation.

Am I affected?

You’re exposed only if both are true:

1. The OS Query content pack is installed at a version from 1.0.0 up to, but not including, 1.0.17.
2. The SSH integration is installed and configured.

Both Cortex XSOAR and Cortex XSIAM are affected. If the OS Query pack isn’t installed, or the SSH integration isn’t configured, the advisory says the issue isn’t exploitable.

To check, look at the installed version of the OS Query pack in your content pack list. The advisory gives no menu path or command for this, so use your platform’s normal content pack management view.

What to do

1. Update the OS Query content pack to 1.0.17 or later from the Cortex XSOAR Marketplace.
2. If you can’t update yet, disable the OSQueryBasicQuery automation. This is Palo Alto Networks’ workaround. Playbooks that call it will stop working until you re-enable it after updating.
3. Review who holds administrator rights in XSOAR and XSIAM. The bug needs a privileged account, so trim the list and check that those accounts are protected with strong authentication.
4. If you suspect misuse, review activity from the SSH integration’s accounts on the remote endpoints (see Detection) and treat any unexplained commands as a possible compromise of those endpoints. Rotate the credentials or keys the SSH integration uses.

Detection

The advisory and CVE record publish no indicators of compromise and no detection guidance. Based on how the bug works, these are reasonable places to look:

• Runs of the OSQueryBasicQuery automation that you don’t recognise, especially by users who don’t normally run it. Check the platform’s audit and war room records.
• Unexpected commands in the shell history and authentication logs of the endpoints that the SSH integration connects to, around the time of those runs.
• Logins to those endpoints by the SSH integration’s account from outside your normal automation windows.

These are suggestions, not vendor-published signatures.