Security researchers and honeypot operators have confirmed active exploitation of three critical vulnerabilities in Fortinet’s FortiSandbox platform, with live attack telemetry capturing real-world exploitation attempts over the past 24 hours. Among the flaws being targeted is CVE-2026-39813, a vulnerability with no previously recorded exploitation history, making these attacks a first-of-their-kind event.

Three Critical CVEs Under Active Attack

Threat intelligence platform Defused has flagged three CVEs currently under active targeting. All three vulnerabilities can be exploited without authentication through a single crafted HTTP request to the /jsonrpc/ API endpoint over port 443. This means any exposed FortiSandbox management interface is at risk — no credentials or prior access required.

CVE-2026-39813 is a path traversal vulnerability (CWE-24) in the FortiSandbox JRPC API. By injecting traversal sequences such as session: "../../tmp/" into API requests, unauthenticated remote attackers can bypass authentication entirely and access sensitive system data — including configuration backups, serial numbers, and version details. This CVE has never before been observed in active exploitation.

CVE-2026-39808 is an OS command injection flaw (CWE-78) in a FortiSandbox API endpoint. A public proof-of-concept exploit has been available since April 2026, weaponizing the jid GET parameter via pipe-chained Unix commands to execute arbitrary commands as root. Attack payloads consistent with this PoC have now been observed in live exploitation attempts.

CVE-2026-25089 is a second OS command injection vulnerability (CWE-78) affecting the FortiSandbox Web UI across versions 5.0.0–5.0.5, 4.4.0–4.4.8, 4.2 all versions, and FortiSandbox Cloud/PaaS deployments. Interestingly, no functional public exploit has been disclosed for this CVE. Observed exploitation attempts appear to be AI-assisted or heuristically generated exploits with faulty logic, suggesting opportunistic actors are probing without a validated working payload.

Affected Versions and Fixes

• CVE-2026-39813: FortiSandbox 4.4.0–4.4.8, 5.0.0–5.0.5 — Fixed in 4.4.9 and 5.0.6+
• CVE-2026-39808: FortiSandbox 4.4.0–4.4.8 — Fixed in 4.4.9+
• CVE-2026-25089: FortiSandbox 4.2 all versions, 4.4.0–4.4.8, 5.0.0–5.0.5, Cloud/PaaS 5.0.4–5.0.5 — Fixed in 4.4.9 and 5.0.6+

Why This Matters: The FortiSandbox Risk Profile

FortiSandbox is a security analysis platform designed to detect advanced threats and malware. Paradoxically, a compromised FortiSandbox can be weaponized to approve malicious files as clean to dependent Fortinet products — turning the very system meant to protect an organization into a backdoor for attackers. Beyond this, a compromised FortiSandbox can also serve as a lateral movement pivot within enterprise networks, granting attackers a trusted foothold deep inside security infrastructure.

The attacker IP observed in active exploitation — 141.11.43[.]175 — is attributed to AS136510, Streamline Servers Pty Ltd (Singapore), and carries a high-interest threat score in threat intelligence platforms.

Indicators of Compromise (IOCs)

• Attacker IP: 141.11.43[.]175 (AS136510 — Streamline Servers, Singapore)
• Target Port: 443 (HTTPS/JRPC API)
• Target Endpoint: /jsonrpc/
• User-Agent observed: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/149.0.0.0 Safari/537.36

Recommended Actions

Organizations running FortiSandbox should immediately update to fixed versions (4.4.9 or 5.0.6+). Until patching is complete, administrators should restrict access to the FortiSandbox management interface by placing it behind a VPN or firewall rules that limit exposure to trusted IP ranges only. Public-facing management interfaces should be taken offline where operationally feasible.

Security teams should also monitor for exploitation attempts against the /jsonrpc/ endpoint and any anomalous outbound connections from FortiSandbox hosts, which could indicate the device has already been compromised and is being used as a pivot point. Given that one of the targeted CVEs (CVE-2026-39813) had never previously been exploited in the wild before this campaign, organizations should not assume that absence of prior exploitation history means lower risk.