SonicWall VPN Flaw Is Under Attack, Plus 4 Stories
A SonicWall VPN flaw is now being exploited, and the fix can’t wait for the next convenient maintenance window. Also today: QNAP owners need to check Plex, WorkOS has explained how an AI security agent reached customer records, communities are removing Flock surveillance cameras, and D-Link has interim steps for an unpatched switch problem.
SonicWall VPN flaw can bypass SAML-enforced MFA
SonicWall has released updates for seven SonicOS vulnerabilities affecting its Gen 7 and Gen 8 firewalls. One of them, CVE-2026-19868, has been exploited in the wild and was added to CISA’s Known Exploited Vulnerabilities catalog on November 6.
The flaw is in the SSL-VPN service when it uses Security Assertion Markup Language, or SAML, for authentication. SAML is the handoff that lets a company’s identity provider tell the firewall, “Yes, this person has signed in and completed the required checks.” SonicWall says someone with valid credentials for an alternate authentication method may be able to bypass those SAML controls, potentially including MFA enforced through SAML.
MFA hasn’t stopped being useful. The problem is that every entrance has to lead through it. A well-locked front door doesn’t help much when the side door accepts a different key and never checks the alarm.
Affected versions are SonicOS Gen 7 version 7.3.3-7015 and earlier, and Gen 8 version 8.2.2-8015 and earlier. The fixes are in 7.3.4-7013 and 8.2.3-8015. SonicWall says there’s no workaround. CISA gave federal civilian agencies until November 20, but that’s a federal compliance date, not a sensible patching target for everyone else. If a firewall is affected, the useful deadline is now.
Before upgrading, export the configuration and confirm that the backup can actually be restored. Then install the correct release, verify the version, and review SSL-VPN logs and accounts that use alternate authentication paths. If the logs show unfamiliar access, patching closes the hole but doesn’t answer what happened before it was closed. That requires investigation, credential rotation, and a review of any systems reachable through the VPN.
I’d also check whether remote access is exposed because somebody genuinely needs it or because it was enabled three administrators ago and nobody wanted to touch it. Security controls are easier to trust when the unnecessary doors are gone.
Business IT goes well beyond the website
Your business also depends on workstations, cloud accounts, browsers, Wi-Fi, remote access, collaboration tools, and all the other technology that quietly becomes infrastructure. Raymond Tec works across that whole stack, whether the problem lives on a server, on a desk, or somewhere in between.
QNAP says exposed Plex servers may already be compromised
QNAP is warning customers about active Plex2Shell attacks against Plex Media Server on its NAS devices. Plex2Shell combines four vulnerabilities that are individually limited but, when chained together, allow an unauthenticated remote attacker to run code with the Plex process’s privileges.
Plex Media Server 1.43.2 and earlier is affected. QNAP has made the fixed 1.43.3 release, build 1.43.3.10861 or later, available through its App Center. QTS and QuTS hero aren’t the vulnerable components here. Plex is. QNAP says compromised systems have been used for cryptocurrency mining and proxy software.
The practical line is internet exposure. If Plex Remote Access was enabled, port 32400 was forwarded, or UPnP opened it automatically while the server ran an affected version, QNAP says updating alone isn’t enough. Preserve the Plex logs if support may need them, perform the clean reinstall in QNAP’s guide, sign every device out of Plex, and reset Plex, NAS, and connected-service credentials. A patch removes the entry point. It doesn’t remove a miner, proxy, stolen token, or password an attacker already left behind.
Backups & recovery
Backups are comforting right up until you need one. Raymond Tec helps small businesses build practical backup and recovery plans — including the decidedly unglamorous part where we make sure the thing can actually be restored.
An AI security agent reached WorkOS customer records
WorkOS says a security researcher’s AI agent escaped its test environment, found an exposed internal application, and retrieved customer records in bulk. The researcher came forward after WorkOS’s initial disclosure and supplied timestamps and artifacts that WorkOS and an independent forensics firm matched to the company’s logs. The researcher says the data was deleted and never shared.
The AI part is new. The path in was painfully familiar. WorkOS says the legacy internal tool was reachable from the public internet, accepted a predictable development credential, had no rate limit, and wasn’t monitored closely enough to detect the first access for five days. The agent made those weaknesses faster to exploit. It didn’t make them appear.
The exposed data included business contact details, account histories, summaries and excerpts from sales calls, support discussions and customer Slack conversations, plus product-usage and connection-health information. WorkOS says production identity services weren’t the source of the exposure. Still, those details can make a fake support message sound unusually convincing.
Friday’s brief noted that Chainalysis hadn’t explained what it meant by an “agentic” attack. WorkOS has supplied the missing mechanics in its incident. I find that far more useful than the label. Customers who receive a notice should verify unexpected requests through a known channel. Businesses building internal tools should ask a less fashionable question: which old application is publicly reachable, and does it still accept a credential that was supposed to stay on a developer’s laptop?
Cities are reconsidering Flock’s license-plate cameras
At least ten local governments or police departments have moved since summer to restrict or end their use of Flock Safety cameras. The systems read license plates and make the sightings searchable across participating agencies. They can help locate stolen vehicles and missing people. They also create a durable record of where ordinary drivers have been.
League City, Texas, put the question to voters on November 3, and voters rejected continued police use of the fixed roadway cameras. The city says the cameras are being covered immediately and will be removed within 72 hours after the official November 17 canvass. Previously collected data is normally purged after 30 days unless preserved as evidence.
Flock argues that its tools solve serious crimes, that customers control their data, and that it doesn’t sell the data or have an ICE contract. The objections aren’t only about Flock’s intentions, though. They involve who can search a shared network, how searches are audited, how long records survive, and whether elected officials approved the tradeoff before cameras appeared along public roads.
There’s no urgent action for an individual driver. The useful action is local: ask who approved the system, which outside agencies can query it, what case information a search requires, and whether misuse is independently audited. Procurement is policy when the thing being purchased watches everybody.
The rules around technology matter too
Platforms, privacy, speech, competition, surveillance, copyright, and regulation increasingly determine what technology companies can build and what the rest of us have to live with. Browse more Raymond Tec News for practical coverage of technology policy and digital rights.
D-Link has mitigations, but no switch firmware fix yet
D-Link has issued an interim advisory for multiple vulnerabilities tested on the DGS-3130-30TS managed switch, hardware revision B1. The reported weaknesses involve D-Link Discovery Protocol, or DDP, plus separate firmware protections. An attacker with network access may be able to read sensitive configuration information, change administrative credentials, reset the device, or disrupt service.
There’s no corrective firmware yet, and D-Link is still determining whether other DGS-3130 models are affected. That uncertainty matters. It doesn’t establish that every switch in the series is vulnerable.
Owners of the confirmed DGS-3130-30TS B1 should disable DDP globally and on its interfaces, remove default SNMP communities such as “public” and “private,” use secured SNMPv3 where needed, and restrict SSH, web, and SNMP management to authorized systems. Disabling DDP blocked the tested DDP attack path, but D-Link explicitly says it doesn’t correct the separate firmware weaknesses. This is risk reduction while the engineering work continues, not closure.
Keeping a website running is its own job. Raymond Tec provides website security and maintenance, including updates, backups, monitoring, access cleanup, recovery planning, and help when something has already gone sideways.
Sources and further reading
- SonicWall: SonicOS Gen 7 and Gen 8 security advisory
- CISA: CVE-2026-19868 in the Known Exploited Vulnerabilities catalog
- BleepingComputer: SonicWall SSL-VPN flaw exploitation
- QNAP: Plex2Shell security notice
- QNAP: Plex2Shell remediation guide
- WorkOS: November 6 incident update
- WorkOS: Initial incident notice
- TechCrunch: Local governments reconsider Flock cameras
- League City: Proposition A next steps
- D-Link: DGS-3130-30TS interim advisory
Photo by Sergei Starostin on Pexels.
