Synology published a notice on 29 October saying threat actors were exploiting a vulnerability it patched in December 2024 to encrypt data on outdated DiskStation Manager builds and demand a ransom 1. The NVD record lists the CISA KEV addition on 18 November 2026 with a due date of 21 November 2026 2, and CISA’s alert the same day names it among four new additions 3. The patch has been available for close to two years, so the question I care about is which devices still answer a request from the internet, and how they got that way.
The bug in the system plugin daemon
Synology describes CVE-2024-10441 as an improper encoding or escaping of output bug in the system plugin daemon, and it lets remote attackers execute arbitrary code 4. The affected builds are DSM before 7.2-64570-4, 7.2.1-69057-6 and 7.2.2-72806-1, plus BeeStation OS before 1.1-65374 4. The advisory lists DSM 7.1 and DSM 6.2 in its affected-product table, with fixed builds at 7.1.1-42962-7 and 6.2.4-25556-8 4, while the CVE-2024-10441 detail text names only the 7.2 branches and BeeStation OS 4. I read both parts of the page before I call a build out of scope.
Synology rates the flaw 9.8 critical on CVSS 3.1, with a vector of AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H 4. That vector describes the attacker: network reachable, no account required, nothing for the victim to click. The advisory’s mitigation field reads None, and the flaw came out of Pwn2Own 2024, with the advisory crediting researchers from DEVCORE, Ryan Emmons and Team Smoking Barrels 4.
The vendor notice and the KEV entry name different things
Synology’s October notice does not name a CVE 1. BleepingComputer asked which flaw was being exploited in the campaign, and a Synology spokesperson pointed back to the November 2024 advisory that covers three: CVE-2024-10441 for code execution, CVE-2024-50629 for limited file reads and CVE-2024-10445 for limited file writes 5. CISA’s entry names CVE-2024-10441 alone 3. The specific exploitation claim is CISA’s, and I write it down that way. For planning the choice costs me nothing, since the 9.8 with code execution is the one the KEV clock is attached to.
The campaign from outside
BleepingComputer reported that victims found a ransom note named read-me3.txt and encrypted files carrying a .L0CK3D extension, and that the malware borrows the Cerber name from the 2016 operation of the same name 5. The same report cites Shadowserver’s dashboard at more than 380,000 internet-exposed Synology NAS devices, with no breakdown of how many had been patched 5. HKCERT’s alert on 2 November carries the guidance that matters more than the version number: keep the NAS management interface off the internet, and disable the services and port forwards that nothing needs 6.
Where the boundary sits
My rule for a while was that anything on a private address was internal. That rule held until I sat down with two lists side by side. The first is the forward table on the router: one row per external port, mapped to an internal address and port. The second is the host list my reverse proxy publishes. A service can appear in the second one with an auth gate in front of it and still show up in the first on a port I opened for something else years ago. The row that matters is the one in the forward table that is not in the proxy list, because nothing in my normal workflow reads it.
A two-year-old patch is the case I use to test the rule. A device that sat unpatched through two years of update cycles had nobody watching it, and the same gap leaves a forward in place.
The reconciliation I run
Four steps. Only the fourth one has surprised me.
- Dump the router’s forward table, disabled rows included, because they show what used to be reachable.
- Dump the proxy’s published hostnames and the backend each one targets.
- Match the two lists, then write down what answers on the far side: the app’s own login, an auth gate in front of the app, or an endpoint that asks nothing.
- Test from outside the network, from a phone on cellular. A test from the LAN proves less than it looks, because split DNS and NAT hairpinning send the request from inside along a path an outside client never takes.
Step four is the one I skip when I am in a hurry, and it is the one that catches a hostname that resolves from the internet while I have it filed as internal.
The proxy forwards the bug too
The reverse proxy earns its place. One certificate store, one log to read when something looks wrong. CrowdSec gives me the other half, a running list of what is being probed. Neither of them changes what happens once the request reaches the backend. A 9.8 that needs no credentials works the same whether the request lands on the app’s own port or on the proxy’s, because the proxy passes the request body through the same way it passes a login form. The proxy still gives me one place to close a path while I check versions, and that beats editing firewall rules per service at the moment a KEV clock starts.
The order I work in
Closing the path comes first, because it does not wait on a vendor. Block the forward, or restrict the management interface to the LAN, and the exposure changes while the patch queue catches up. Version check second, against the fixed builds in the advisory. Compromise assessment third, and that step is not optional for a device that sat exposed: CISA’s alert sets an expectation that agencies check whether an attacker got in before the patch went on 3. Restore fourth, from a copy the NAS cannot reach over the network. HKCERT’s guidance covers the reason, since a backup job the affected device can write to carries the encryption into the backup set 6.
The notes I keep afterwards
Two things go into the notes file after a week like this. The first is the forward table, copied in with the date I read it, so the next comparison has a baseline. The second is the class of device rather than the model: an appliance with a management interface, a vendor update path and a person who bought it. That shape produces a missed patch and a forgotten forward in the same device. A KEV entry on that shape sends the version question to whoever owns it, and the exposure question stays on my list, because that half I can answer without their help.
The forward table is short enough to read in one sitting, and nothing on it needs a vendor to fix.
-
Synology. (2026, October 29). Synology urges users to update DSM following ransomware attacks on outdated DSM versions. Synology Inc. https://www.synology.com/en-global/company/news/article/C3RB3R ↩︎ ↩︎
-
National Vulnerability Database. (2025, March 19; last modified 2026, November 19). CVE-2024-10441 detail. National Institute of Standards and Technology. https://nvd.nist.gov/vuln/detail/CVE-2024-10441 ↩︎
-
Cybersecurity and Infrastructure Security Agency. (2026, November 18). CISA adds four known exploited vulnerabilities to catalog. CISA. https://www.cisa.gov/news-events/alerts/2026/11/18/cisa-adds-four-known-exploited-vulnerabilities-catalog ↩︎ ↩︎ ↩︎
-
Synology. (2024, November 5; last updated 2025, March 28). Synology-SA-24:20 DSM (PWN2OWN 2024). Synology Inc. https://www.synology.com/en-global/security/advisory/Synology_SA_24_20 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
-
Gatlan, S. (2026, November 2; updated 2026, November 3). Synology warns of C3RB3R ransomware attacks targeting NAS devices. BleepingComputer. https://www.bleepingcomputer.com/news/security/synology-warns-customers-of-ongoing-c3rb3r-ransomware-attacks/ ↩︎ ↩︎ ↩︎
-
Hong Kong Computer Emergency Response Team Coordination Centre. (2026, November 2). Ransomware alert: Ransomware attacks targeting Synology NAS devices. HKCERT. https://www.hkcert.org/security-bulletin/security-alert-ransomware-attacks-targeting-synology-nas-devices_20261102 ↩︎ ↩︎