Attackers are actively exploiting a set of vulnerabilities in Fortinet FortiSandbox, the appliance many organizations rely on to detonate and analyze suspicious files, and turning it into a route to seize control of the very system meant to catch malware. The exploitation was flagged in threat intelligence published by Check Point Research.
What's affected
Three flaws, tracked as CVE-2026-39813, CVE-2026-39808, and CVE-2026-25089, can be reached through unauthenticated API requests, meaning an attacker does not need a valid login to trigger them. Chained together they allow path traversal (reading or writing files outside the intended directory) and command execution with root privileges, the highest level of access on the device.
Why it matters
FortiSandbox sits at the heart of a defender's malware analysis and policy enforcement workflow. An attacker who takes it over does not just gain another server; they gain a trusted vantage point that can be used to suppress detections, tamper with the verdicts rendered on malicious samples, and pivot deeper into the security stack. Because the flaws need no authentication, any FortiSandbox management or API interface exposed to untrusted networks is at immediate risk.
What you should do
- Apply Fortinet's fixes for CVE-2026-39813, CVE-2026-39808, and CVE-2026-25089 as a priority, and consult Fortinet's advisories to confirm affected versions.
- Remove FortiSandbox administrative and API interfaces from internet exposure and restrict them to trusted management networks.
- Hunt for signs of compromise such as unexpected files, new accounts, or anomalous API calls, and treat any takeover as a foothold that may extend to connected systems.
The flaws add to a busy stretch of actively exploited enterprise software; see our coverage of a critical pre authentication flaw in Splunk Enterprise and a separate credential theft campaign against Fortinet and Sophos devices.
This briefing is provided by IntelFusions for informational and defensive purposes only. It is based on sources assessed to be reliable at the time of writing, and analytic judgments carry the confidence levels indicated. Indicators of compromise are defanged; re-arm them only in controlled environments. IntelFusions is not affiliated with the organizations named and makes no warranty as to completeness or accuracy.