Safeguard
Vulnerability Analysis•

CVE-2025-66261: Critical Command Injection in Mozart FM Restore

CVE-2025-66261 is unauthenticated command injection in restore_settings.php on Mozart FM. No fix is named, so isolate the interface and ask the vendor.

Safeguard Security Team
4 min read

CVE-2025-66261 is a critical (CVSS 9.8) unauthenticated OS command injection in the Mozart FM Transmitter. The restore_settings.php page decodes a name request parameter and passes it into exec() without validation or escaping. The record names no fix, so isolate the management interface and contact the vendor.

What's affected

  • Product: Mozart FM Transmitter from DB Electronica Telecomunicazioni S.p.A. (DB Broadcast), web management interface. The record lists models 30, 50, 100, 300, 500, 1000, 2000, 3000, 3500, 6000 and 7000.
  • Component: /var/tdf/restore_settings.php and its name parameter, which is URL decoded and passed to exec().
  • Fixed: neither the record nor the researcher write-up we read names a fixed firmware version or a vendor advisory. Contact the vendor for patch status before assuming one exists.
  • Severity: the record scores it 9.8 (critical) with the vector CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H. Both the record and the write-up treat this as one of the most serious flaws in the set.

How to check if you're exposed

This is a broadcast appliance, so there is no package manager to query. Work from the device and the network.

  1. Build a list of every Mozart unit you run, with its model and the web interface version string. The researcher write-up refers to the web management interface version WEBMOZZI-00287.
  2. Find out whether the management web interface of each unit is reachable from the internet, from a partner network or from a flat internal network. Check firewall rules and any remote access gateway.
  3. Look at the web server logs on a device you can inspect, or at the logs of any proxy in front of it, for requests to the PHP file named in this CVE from addresses you do not recognize.

If the interface is reachable by people outside the small group that needs it, treat the device as exposed.

What to do now

  1. Take the management interface off the internet and off any shared network. Put it behind a VPN or a jump host, with an allow list of administrator addresses.
  2. Ask DB Broadcast for the patch status of this CVE and for a firmware version that fixes it. Record the answer with your asset notes.
  3. If you have logging at a proxy, alert on requests to restore_settings.php from any address outside your administrator list.
  4. Watch the device for unexpected files, new accounts, changed settings or reboots, and keep a known good copy of its configuration to compare against.
  5. Add the rest of your internet facing software to an inventory. Safeguard cannot patch a transmitter, but a free in-browser scan at gold.safeguard.sh/scan and a proper asset inventory help you find the other internet facing software you run, and reachability analysis helps you rank what to fix first.

Why it matters

The weakness is CWE-78, OS command injection. The record notes the value goes through URL decoding before it reaches the shell, which can let characters that a simple filter would catch reach the command anyway. Decoding after validation, or validating the wrong form of the data, is a frequent cause of bypasses.

Together with the start_upgrade.php flaw, this shows more than one page in the same interface passing request data to a shell. Fixing one would not close the others, so the only safe stance is to limit who can reach the interface at all and press the vendor for a full review. The record does not mark this as exploited in the wild and it is not on the CISA KEV list. The advisory trail is thin, since the main public source is a researcher write-up rather than a vendor bulletin, so confirm facts with the vendor. Appliances of this kind sit for years without updates, which is why network isolation is the control you can apply today.

Quick check

Why does URL decoding matter here?

If input is decoded after any checks, characters can reach the command that the check was meant to block.

Is this the only command injection in the product?

No. The record set includes others, including start_upgrade.php and main_ok.php.

Sources

Last reviewed: 18 October 2026.

Never miss an update

Weekly insights on software supply chain security, delivered to your inbox.

Self-healing security runs on Safeguard.

Your first fix PR is minutes away.

No sales call required, even your agent can complete the purchase over MCP.