Critical FortiSandbox flaw lets unauthenticated attackers run commands

Fortinet has disclosed a critical FortiSandbox vulnerability that can let remote, unauthenticated attackers execute unauthorized commands through the web interface. The flaw is tracked as CVE-2026-25089 and affects FortiSandbox, FortiSandbox Cloud, and FortiSandbox PaaS deployments.

The issue appears in Fortinet advisory FG-IR-26-141, published on June 9, 2026. Fortinet says the vulnerability involves improper neutralization of special elements used in an OS command, also known as OS command injection.

The risk is serious because exploitation does not require authentication. A remote attacker can send crafted HTTP requests to the FortiSandbox web UI and execute unauthorized commands on the affected system.

What CVE-2026-25089 affects

CVE-2026-25089 affects FortiSandbox systems running vulnerable 4.4 and 5.0 builds, plus specific FortiSandbox Cloud and FortiSandbox PaaS versions. Fortinet’s advisory lists the issue as Critical and says it was discovered internally by Adham El Karn of the Fortinet Product Security team.

The NVD entry for CVE-2026-25089 also describes the flaw as an OS command injection issue that may allow an unauthenticated attacker to execute unauthorized commands through specially crafted HTTP requests.

Fortinet says there were no known reports of exploitation when the advisory was published. Even so, unauthenticated command execution flaws in internet-accessible security appliances often draw quick attention from attackers after disclosure.

Product	Affected versions	Fixed version
FortiSandbox 5.0	5.0.0 through 5.0.5	Upgrade to 5.0.6 or above
FortiSandbox 4.4	4.4.0 through 4.4.8	Upgrade to 4.4.9 or above
FortiSandbox Cloud 5.0	5.0.4 through 5.0.5	Upgrade to 5.0.6 or above
FortiSandbox PaaS 5.0	5.0.4 through 5.0.5	Upgrade to 5.0.6 or above

Why the FortiSandbox command injection bug is critical

FortiSandbox is used to analyze suspicious files, URLs, and malware behavior. Fortinet describes FortiSandbox as an AI-powered sandboxing platform for detecting emerging, sophisticated, and zero-day threats.

That role makes the system valuable to defenders, but it also makes it valuable to attackers. If a FortiSandbox instance is compromised, attackers may gain access to a security tool that processes suspicious files, interacts with other parts of the security stack, and supports threat detection workflows.

The vulnerability affects the web UI component. Since exploitation can happen through HTTP requests and does not require a valid account, exposed management interfaces create the highest risk.

• The attack type is unauthenticated.
• The affected component is the FortiSandbox GUI.
• The impact is unauthorized command execution.
• The weakness class is CWE-78, OS command injection.
• Fortinet marked the issue as not known to be exploited at publication.

Fixed and unaffected FortiSandbox versions

Fortinet says FortiSandbox 5.2 is not affected. FortiSandbox Cloud 4.4, FortiSandbox Cloud 5.2, FortiSandbox PaaS 4.4, FortiSandbox PaaS 5.2, and FortiSandbox PaaS 23.4 are also listed as not affected.

Organizations running affected on-premises FortiSandbox builds should upgrade to 5.0.6 or 4.4.9 or later. Cloud and PaaS customers running affected 5.0 builds should move to 5.0.6 or above.

The FortiGuard PSIRT advisory does not list a workaround section, so patching should be the main remediation path. Restricting web UI access to trusted management networks can reduce exposure, but it should not replace the update.

Version line	Status
FortiSandbox 5.2	Not affected
FortiSandbox 5.0.0 through 5.0.5	Affected
FortiSandbox 5.0.6 and later	Fixed
FortiSandbox 4.4.0 through 4.4.8	Affected
FortiSandbox 4.4.9 and later	Fixed

CVSS score discrepancy

There is one scoring detail administrators should note. Fortinet’s advisory page lists a CVSSv3 score of 9.1 for CVE-2026-25089, while the NVD vulnerability record currently shows a Fortinet CNA CVSS 3.1 base score of 9.8 with network attack vector, low complexity, no privileges required, and no user interaction.

Both scores place the flaw in the Critical range. The operational priority does not change: exposed FortiSandbox web interfaces running affected versions should be patched quickly.

Security teams should also review whether the FortiSandbox web UI is reachable from untrusted networks. Management interfaces for security appliances should not be directly exposed to the internet unless there is a strong, documented business need and layered access control.

What administrators should do now

Administrators should first identify all FortiSandbox, FortiSandbox Cloud, and FortiSandbox PaaS instances in their environment. They should then compare each version against Fortinet’s affected-version table and schedule upgrades for any affected builds.

Organizations that use Fortinet’s sandboxing platform as part of malware analysis, email security, or network defense should treat the appliance as a high-value asset. Logs from the web UI, reverse proxies, VPN gateways, and management networks should be reviewed for suspicious HTTP requests.

• Upgrade FortiSandbox 5.0.0 through 5.0.5 to 5.0.6 or later.
• Upgrade FortiSandbox 4.4.0 through 4.4.8 to 4.4.9 or later.
• Upgrade affected FortiSandbox Cloud and PaaS 5.0 deployments to 5.0.6 or later.
• Restrict FortiSandbox web UI access to trusted administrator networks.
• Review logs for unexpected HTTP requests to the web interface.
• Check firewall and VPN rules for unnecessary public exposure.
• Monitor FortiSandbox systems for unusual processes, command activity, or configuration changes.

CVE-2026-25089 gives attackers a direct path to command execution if they can reach an affected web interface. Fortinet has not reported active exploitation, but the lack of authentication makes this a patch-now issue for any organization running vulnerable FortiSandbox builds.

FAQ