Fortinet FortiSandbox Vulnerability Allows Critical Command Execution
A critical OS command injection vulnerability (CVE-2026-25089) in Fortinet's FortiSandbox web UI allows unauthenticated attackers to execute arbitrary commands, potentially leading to full system compromise.
Fortinet has issued a critical security advisory for its FortiSandbox product line, detailing a vulnerability that permits unauthenticated remote attackers to execute arbitrary OS commands. The flaw, identified as CVE-2026-25089, carries a CVSSv3 score of 9.1, classifying it as Critical. This vulnerability affects multiple versions of FortiSandbox, FortiSandbox Cloud, and FortiSandbox PaaS deployments.
The root cause of the vulnerability is an improper neutralization of special elements used in an OS command, a common weakness known as OS command injection (CWE-78). This flaw is present within the FortiSandbox Web UI. By crafting and sending specific HTTP requests, a remote attacker who has not authenticated can exploit this weakness to inject and execute unauthorized commands on the underlying system.
The severity of CVE-2026-25089 is amplified by its low attack complexity and the lack of an authentication requirement. This means that threat actors can exploit it relatively easily, and a successful compromise can lead to a complete loss of confidentiality, integrity, and availability for the affected system. The vulnerability was discovered and reported internally by Adham El Karn from Fortinet's Product Security team.
Fortinet has provided specific affected versions and their corresponding fixes. For FortiSandbox versions 5.0.0 through 5.0.5, an upgrade to version 5.0.6 or later is required. Similarly, FortiSandbox versions 4.4.0 through 4.4.8 must be upgraded to 4.4.9 or later. FortiSandbox Cloud and PaaS deployments running versions 5.0.4 through 5.0.5 also need to be updated to 5.0.6 or above. Notably, versions 5.2, 4.4, and 5.2 of FortiSandbox Cloud, and versions 4.4, 5.2, and 23.4 of FortiSandbox PaaS are not impacted.
While Fortinet has not reported any instances of active exploitation in the wild, the unauthenticated nature of this vulnerability makes it a prime target for malicious actors. FortiSandbox is a critical component in many enterprise security infrastructures, used for malware analysis and threat detection. A compromise of this system could provide attackers with a strategic advantage, potentially allowing them to bypass or disable an organization's defenses.
Fortinet strongly advises security teams to take immediate action. The primary recommendation is to upgrade all affected FortiSandbox installations to the patched versions as soon as possible. As a temporary mitigation measure, organizations can restrict access to the web UI to only trusted IP address ranges. Additionally, monitoring system logs for any anomalous HTTP requests directed at the FortiSandbox web interface is recommended.
Given the critical nature of this vulnerability and the ease with which it can be exploited, organizations running any of the affected versions should prioritize patching. This situation underscores the importance of maintaining up-to-date security software and promptly applying vendor-provided patches to protect against sophisticated cyber threats.
In addition to the critical FortiSandbox vulnerability, Fortinet has also released patches for medium-severity flaws affecting FortiOS, FortiProxy, and FortiPortal. These lesser vulnerabilities could allow authenticated users to execute scripts or disclose sensitive network configuration data, though Fortinet has not reported any in-the-wild exploitation for any of these issues.
Threat intelligence firm Defused has now confirmed active exploitation of three critical FortiSandbox vulnerabilities, including CVE-2026-39813, CVE-2026-39808, and CVE-2026-25089, with attacks observed within the past 24 hours. The flaws allow unauthenticated remote code execution and privilege escalation via low-complexity command injection. Fortinet released patches on April 14, and organizations are urged to upgrade immediately to block incoming attacks.
Threat intelligence firm Defused Cyber reports that three vulnerabilities in FortiSandbox are now being actively exploited, including CVE-2026-25089 (patched last week), CVE-2026-39813, and CVE-2026-39808 (both patched in April 2026). The exploit for CVE-2026-25089 appears to be AI-assisted but is faulty, with no working public exploit yet available.
Threat intelligence firm Defused reported on Monday that attackers are actively exploiting all three vulnerabilities, including the previously disclosed CVE-2026-25089, CVE-2026-39813, and CVE-2026-39808. Notably, the exploit for CVE-2026-25089 appears to be "vibecoded" and likely faulty, suggesting AI-assisted development may be lowering the barrier for attackers to target even historically less common Fortinet products. Fortinet has released patches for all three flaws, though the vendor has not yet confirmed in-the-wild exploitation.
New telemetry confirms that all three CVEs — including CVE-2026-39813 (path traversal) and CVE-2026-39808 (OS command injection with a public PoC) — are now being actively exploited in the wild, with attacker IP 141.11.43.175 (AS136510, Singapore) targeting the /jsonrpc/ endpoint on port 443. The article also notes that CVE-2026-25089 exploitation attempts appear to be "vibecoded" (AI-assisted) and lack a validated working payload, while a compromised FortiSandbox can be used to approve malicious files as clean or enable lateral movement.
VulnCheck and Defused researchers observed active exploitation of the April-patched vulnerabilities CVE-2026-39808 (OS-command injection) and CVE-2026-39813 (path traversal) starting June 9. Over a six-day period, 49 exploitation events from 11 distinct IPs were recorded, alongside attempts targeting the newly patched CVE-2026-25089. Post-exploitation activity—verification and reconnaissance—precedes a likely heavier wave of attacks, and researchers warn that compromising FortiSandbox could grant attackers elevated access within security-sensitive environments.