Zimbra Under Active Attack as Hackers Exploit a Critical Command-Execution Flaw While 14,000+ Dahua Cameras Fall in CameraSwarm + Video

Listen to this Post

Featured Image

A Dangerous Day for Internet-Exposed Infrastructure

Cybersecurity is increasingly being defined by attacks that do not need to begin with a sophisticated zero-day or a massive ransomware deployment. Sometimes, the most dangerous campaigns start with something much simpler: an exposed service, an outdated device, or a security feature that was never expected to become an attack path.

Two developments highlighted on August 20, 2026, demonstrate exactly how quickly that risk is expanding. CERT Polska has warned that attackers are actively exploiting CVE-2026-73570, a command-injection vulnerability in Zimbra Collaboration. At the same time, researchers have uncovered Operation CameraSwarm, a campaign in which more than 14,000 Dahua cameras were compromised across Ukraine and Russia.

These incidents involve very different technologies, but they reveal the same uncomfortable reality: attackers are increasingly targeting infrastructure that organizations often treat as background technology.

Email collaboration servers sit at the center of corporate communications, while IP cameras sit inside offices, warehouses, homes, factories, and public facilities. Both can become powerful footholds when they are exposed to the internet and left insufficiently protected.

The Zimbra Vulnerability Is Being Actively Exploited

CERT Polska reported on August 17 that CVE-2026-73570 is being actively exploited against Zimbra Collaboration installations. The vulnerability is an OS command-injection flaw that can allow an unauthenticated attacker to execute arbitrary shell commands with the privileges of the zimbra user.

The vulnerability specifically affects installations where SNMP notifications are enabled through the snmp_notify parameter and the swatchdog service is running. CERT Polska says the latter is enabled by default.

That combination makes the issue particularly concerning because exploitation does not require the attacker to first authenticate to the Zimbra environment.

Zimbra 10.1.20 Contains the Fix

Zimbra lists CVE-2026-73570 among its security advisories and identifies version 10.1.20 as the release containing the fix. Independent vulnerability records identify the issue as an unauthenticated remote-code-execution vulnerability and associate it with CWE-78, the category covering OS command injection.

The currently reported CVSS score is 8.9, placing the vulnerability in the high-severity range.

For administrators, the message is therefore straightforward: systems running affected versions should not be treated as safe simply because no suspicious activity has yet been observed.

Attackers Can Execute Commands as the Zimbra User

The distinction between remote access and command execution matters enormously here. An attacker who can execute operating-system commands has moved beyond merely interacting with the application.

Depending on the server’s configuration and surrounding defenses, command execution could allow attackers to inspect files, establish persistence, deploy additional tools, search for credentials, manipulate application components, or attempt to move deeper into an organization’s network.

CERT Polska specifically warns that exploitation can result in arbitrary shell commands being executed with the privileges of the zimbra account.

That makes this vulnerability a serious operational concern rather than a theoretical software defect.

CERT Polska Provides Hunting Guidance

CERT Polska has also provided administrators with specific indicators to investigate. Organizations are advised to review Zimbra logs for unusual service-status changes and inspect files created during the previous 30 days by the zimbra user.

The locations highlighted by CERT Polska include /opt/zimbra/jetty/webapps/, /opt/zimbra/jetty_base/webapps/, and /tmp/.

This is particularly important because patching an exploited system does not automatically remove an attacker’s persistence.

Patching Alone May Not Be Enough

If an attacker has already executed commands on a Zimbra server, administrators should approach the system as potentially compromised rather than assuming that installing the patch closes the incident.

The appropriate response depends on the environment, but investigation should include reviewing authentication and application logs, checking unexpected files and processes, examining outbound connections, reviewing privileged accounts, and determining whether credentials stored or accessible from the server may have been exposed.

A patched server can still contain malicious files or altered configurations left behind during an earlier intrusion.

Operation CameraSwarm Reveals a Different Kind of Infrastructure Crisis

While the Zimbra campaign targets enterprise software, Operation CameraSwarm demonstrates how vulnerable physical-security devices can become part of a large-scale cyber operation.

Researchers at Hunt.io reconstructed a campaign involving more than 14,530 Dahua cameras and network video recorders between June 17 and July 22, 2026. The compromises were concentrated in Ukraine and Russia, although the broader lesson extends far beyond those countries.

The original report circulating on X summarized the campaign as involving more than 14,000 cameras and described brute-force attacks, authentication bypasses, and persistent access mechanisms.

The Attack Was Larger Than the Initial Social Post Suggested

The figure of 14,000 cameras is accurate as a rounded description, but subsequent reporting provides a more precise number: 14,530+ devices were compromised during the campaign period.

That scale makes CameraSwarm particularly significant.

This was not simply an attacker compromising a handful of poorly secured cameras. Researchers reconstructed a systematic operation involving automated tooling, target lists, logs, exploitation components, and persistent access mechanisms.

The Attacker Left Behind a Remarkable Digital Trail

One of the strangest aspects of Operation CameraSwarm is how researchers obtained visibility into the campaign.

The operator accidentally exposed a working directory containing approximately 407 MB of data and 2,616 files across 234 directories. Researchers were able to reconstruct portions of the operation from tooling, logs, shell history, and other operational material.

In cybersecurity, attackers often spend enormous effort hiding their infrastructure.

Here, the operational environment itself became evidence.

Thousands of Cameras Were Reached Through Multiple Paths

The campaign did not depend on a single vulnerability.

Researchers identified several exploitation routes, including credential brute forcing, authentication-bypass vulnerabilities, and abuse of Dahua’s peer-to-peer relay infrastructure.

That multi-path approach is important because it illustrates how attackers can adapt when one method fails.

If a device resists password attacks, another exploitation path can be attempted. If direct access is unavailable, the attacker’s toolkit can potentially exploit a cloud-assisted relay mechanism.

Old Vulnerabilities Can Remain Extremely Dangerous

One of the most uncomfortable lessons from CameraSwarm is that attackers do not necessarily need newly discovered vulnerabilities.

Researchers reported exploitation involving CVE-2021-33044 and CVE-2021-33045, vulnerabilities that are several years old.

The continued success of older vulnerabilities demonstrates the enormous security debt created by devices that remain deployed for years without consistent firmware management.

A vulnerability can be old in the security community and still be new to the organization that has never patched it.

Persistent Backdoors Raise the Stakes

Researchers identified approximately 1,923 cameras carrying a persistent backdoor account installed through RPC. The account was stored separately from the administrator password, meaning changing the normal password did not necessarily remove the attacker’s access.

Reports further indicate that the backdoor could survive a factory reset on many affected firmware versions.

That detail transforms the incident from a simple unauthorized-login problem into a device-integrity problem.

Factory Reset Cannot Always Be Treated as a Clean Slate

For consumers and organizations alike, factory resetting a compromised device is often considered the ultimate recovery mechanism.

CameraSwarm demonstrates why that assumption can be dangerous.

If malicious access survives the reset process, defenders may believe they have recovered the device while an attacker retains access. That creates a particularly nasty cycle in which compromised hardware can repeatedly return to the organization’s network.

Firmware updates and vendor-provided remediation therefore become much more important than simply changing a password.

Cameras Are More Than Cameras

The security community has spent years warning that IoT devices can become botnet infrastructure.

But surveillance cameras present a unique risk because they combine computing capability, network connectivity, sensitive physical locations, and access to visual information.

A compromised camera can potentially become a platform for surveillance, credential harvesting, network reconnaissance, lateral movement, or participation in broader malicious infrastructure.

Even when the attacker does not steal video, controlling the device itself can create security and operational risks.

CameraSwarm Shows Why Edge Devices Belong in the Security Model

Many organizations still mentally separate their cameras, printers, access-control systems, sensors, and other connected appliances from their cybersecurity architecture.

That separation is becoming increasingly difficult to justify.

A camera has an IP address.

It has firmware.

It communicates over a network.

It may have credentials.

It may expose services.

It may have vulnerabilities.

From an

The Two Incidents Share the Same Fundamental Weakness

Zimbra and Dahua are not the same technology, and these incidents should not be treated as a single campaign.

But they expose the same structural problem: internet-connected infrastructure is often maintained differently from the systems security teams watch most closely.

Email servers may receive regular attention because they are obviously critical.

Cameras and other IoT systems may remain forgotten for years.

Attackers know this.

The Real Target Is Often the Device Nobody Is Watching

Security teams tend to prioritize laptops, servers, cloud accounts, and identity systems.

Those remain important, but attackers increasingly have reasons to look elsewhere.

An overlooked appliance can provide a foothold.

A vulnerable management interface can expose an internal environment.

A forgotten service can provide persistence.

An old camera can become an attacker-controlled node.

The weakest device can sometimes determine the security of the entire environment.

What Organizations Should Learn From the Zimbra Campaign

The immediate lesson for Zimbra administrators is to identify affected systems, determine whether the vulnerable configuration exists, upgrade to a fixed release, and investigate for evidence of exploitation.

Because CERT Polska has specifically described ongoing exploitation, organizations should prioritize this work rather than placing the vulnerability into a normal long-term patch queue.

The presence of active exploitation changes the risk calculation.

What Organizations Should Learn From CameraSwarm

For Dahua administrators, the lesson is broader than installing a firmware update.

Organizations should maintain an inventory of every connected camera and recorder, identify firmware versions, remove unnecessary internet exposure, restrict management access, rotate potentially compromised credentials, and investigate devices that show unexpected behavior.

Where compromise is suspected, defenders should consider the possibility that normal password changes may not be sufficient.

Network Segmentation Can Limit the Damage

A surveillance camera should not have unrestricted access to the same network resources as an organization’s finance systems, identity infrastructure, databases, and employee endpoints.

Network segmentation can reduce the blast radius when an edge device is compromised.

Even if an attacker successfully takes control of a camera, segmentation can prevent that camera from becoming a convenient bridge into more valuable systems.

Outbound Traffic Deserves More Attention

Organizations often focus on inbound attacks because that is where exploitation begins.

Compromised IoT devices make outbound monitoring equally important.

A camera suddenly communicating with unfamiliar infrastructure, making unusual connections, generating abnormal traffic volumes, or contacting systems it has never previously needed may provide an important signal.

Behavioral monitoring can therefore complement traditional vulnerability management.

Deep Analysis: The Bigger Cybersecurity Pattern

Attackers Are Industrializing Exploitation

Both incidents reflect a broader shift toward automation.

A human attacker manually compromising one server or camera at a time is inefficient.

Automated scanners, exploit frameworks, credential attacks, and specialized tooling allow attackers to turn thousands of potential targets into a manageable dataset.

That changes the economics of cybercrime.

Scale Is Becoming More Important Than Novelty

CameraSwarm did not need an unprecedented zero-day to create a major impact.

The operation demonstrates that known vulnerabilities can remain highly valuable when millions of vulnerable devices continue to exist.

This is one reason vulnerability age should never be confused with vulnerability relevance.

Internet Exposure Multiplies Risk

A vulnerable system isolated behind multiple layers of security is one problem.

A vulnerable system directly reachable from the internet is another.

The combination of outdated software and unnecessary internet exposure is particularly dangerous because automated attackers can discover and test targets continuously.

Security Teams Need Better Asset Visibility

You cannot patch what you do not know exists.

You cannot monitor what is absent from your inventory.

You cannot isolate a device that has never been categorized.

The first step in defending modern infrastructure is therefore visibility.

IoT Inventory Should Be Treated as Security Inventory

Cameras, recorders, sensors, access-control systems, network appliances, and other connected devices should be included in asset-management programs.

They should have owners.

They should have firmware records.

They should have documented network locations.

They should have lifecycle policies.

Vulnerability Management Must Include Physical Infrastructure

Traditional vulnerability programs often focus on operating systems and applications.

That model is no longer sufficient.

Modern organizations increasingly depend on embedded software and specialized appliances, making firmware vulnerabilities part of the enterprise security equation.

Persistence Is More Dangerous Than Temporary Access

An attacker who gains temporary access is dangerous.

An attacker who can maintain access after administrators believe they have recovered the device is significantly more dangerous.

The CameraSwarm backdoor findings highlight this distinction.

Recovery Must Be Verified, Not Assumed

After remediation, defenders should verify that malicious accounts, processes, files, configurations, and communication channels have actually disappeared.

A successful recovery process should produce evidence.

Simply rebooting or resetting a device is not evidence of clean recovery.

Zimbra Shows Why Application Features Can Become Attack Surfaces

The Zimbra vulnerability is particularly interesting because it involves SNMP notification functionality.

Security teams may think of monitoring features as defensive components.

Yet any feature that processes external input can potentially become an attack surface.

Monitoring infrastructure must therefore be secured with the same seriousness as user-facing functionality.

Complexity Creates Hidden Attack Paths

Modern enterprise applications contain numerous interconnected components.

A feature that appears unrelated to email functionality can nevertheless interact with system services, monitoring processes, parsers, and operating-system commands.

Attackers specialize in finding those unexpected relationships.

Default Services Deserve Scrutiny

CERT

Defaults are convenient for administrators.

They are also attractive to attackers because predictable configurations can be targeted at scale.

Security Configuration Should Follow Business Need

If a service is not required, disabling it can reduce attack surface.

If it is required, administrators should understand exactly what it exposes and how it processes untrusted input.

Security should not be based on blindly trusting defaults.

Active Exploitation Changes the Priority

A vulnerability that has merely been disclosed deserves attention.

A vulnerability that is being actively exploited deserves immediate attention.

CERT Polska explicitly describes CVE-2026-73570 as actively exploited.

That distinction should drive emergency patching decisions.

Logs Become Critical After Disclosure

Once a vulnerability enters active exploitation, organizations should not limit themselves to patch management.

They should also hunt for evidence that exploitation already happened.

The Zimbra

Threat Intelligence Must Become Operational

Knowing that attackers are targeting a vulnerability is useful.

Knowing exactly what logs to inspect, what files to examine, what services to monitor, and what network behavior to investigate is much more useful.

Threat intelligence has its greatest value when it becomes a concrete defensive action.

Old Vulnerabilities Should Be Reassessed

CameraSwarm demonstrates that years-old vulnerabilities can still be operationally valuable to attackers.

Organizations should periodically reassess old vulnerabilities based on current exploitation activity rather than assuming that age makes them irrelevant.

IoT Security Is Now Enterprise Security

The boundary between information technology and physical infrastructure continues to disappear.

A compromised camera can become a network foothold.

A compromised printer can expose documents.

A compromised access-control system can affect physical security.

A compromised sensor can disrupt operations.

Attackers Do Not Care Which Department Owns the Device

Security teams may classify a camera as a facilities problem.

Facilities may classify it as an IT problem.

The attacker simply sees an exposed machine.

That ownership gap can create exactly the opportunity adversaries need.

The Cloud Does Not Eliminate Device Risk

Cloud-connected cameras and appliances can make deployment easier.

They can also introduce additional infrastructure, APIs, relay services, authentication systems, and trust relationships.

Convenience features must therefore be assessed from an adversarial perspective.

P2P Infrastructure Deserves Particular Attention

CameraSwarm’s use of Dahua’s relay ecosystem is a reminder that security analysis must consider not only the device itself but also the services connecting it to the outside world.

A device may appear protected because it has no obvious public-facing management interface while still being reachable through a vendor-supported connectivity mechanism.

The Security Model Must Follow the Full Data Path

Defenders should ask how a device communicates with users, vendors, cloud platforms, relay services, update servers, and other infrastructure.

Every connection creates another relationship that could potentially be abused.

Attackers Are Looking for Trust

The most valuable security weakness is not always a software bug.

Sometimes it is trust.

A trusted relay.

A trusted service.

A default account.

A trusted internal network.

A device assumed to be harmless.

Attackers repeatedly turn those assumptions into access.

The Human Factor Still Matters

The CameraSwarm

Operational security failures can reveal infrastructure, tooling, usernames, logs, timestamps, target lists, and other intelligence.

Defenders should therefore monitor not only malware but also attacker infrastructure and exposed operational artifacts.

Cybersecurity Is Becoming a Race Between Automation Systems

Attackers are automating discovery, exploitation, credential attacks, persistence, and reconnaissance.

Defenders need automation on their side as well.

Automated asset discovery, patch prioritization, anomaly detection, isolation, and threat hunting can significantly reduce the time between exploitation and response.

The Biggest Risk Is the Forgotten System

The Zimbra incident is a warning for administrators of business-critical servers.

CameraSwarm is a warning for organizations responsible for thousands of connected devices.

Together, they point toward the same conclusion: the forgotten system is becoming one of the most dangerous systems in an organization’s environment.

What Undercode Say:

The Attack Surface Is Expanding Faster Than Most Organizations Can Track

The most important takeaway is not simply that Zimbra has a critical vulnerability or that thousands of Dahua cameras were compromised. The bigger story is that the modern attack surface is expanding into every device that communicates with a network.

Patch Management Cannot Operate in Isolation

Organizations need vulnerability management connected to asset inventory, threat intelligence, network visibility, identity security, and incident response. A patch without visibility into prior exploitation can leave an organization believing it is protected when the attacker is already inside.

Active Exploitation Should Trigger Emergency Procedures

CVE-2026-73570 deserves urgent treatment because CERT Polska has confirmed active exploitation rather than merely warning about a theoretical possibility.

IoT Devices Need Enterprise-Level Protection

CameraSwarm shows that surveillance equipment can become an important component of an attacker’s infrastructure. Organizations should stop treating cameras and similar devices as harmless peripherals.

Persistence Changes the Entire Recovery Process

The discovery of persistent access mechanisms on compromised Dahua devices is especially concerning because conventional password resets may not completely eliminate the threat.

Older Vulnerabilities Can Still Produce New Disasters

The use of vulnerabilities dating back to 2021 demonstrates that organizations cannot rely on vulnerability age as a measure of practical danger.

Security Teams Need to Monitor the Edge

Endpoints and servers remain essential monitoring targets, but connected cameras, appliances, and embedded devices must increasingly be included in detection and response programs.

Segmentation Should Be Standard

An IoT device should not have unrestricted access to critical enterprise systems. Segmentation can turn a successful device compromise into a contained incident rather than a company-wide breach.

Incident Response Must Assume Persistence

When exploitation is confirmed, defenders should investigate what attackers may have changed, created, downloaded, or installed before declaring the system clean.

The Best Defense Is Visibility

Organizations that know exactly what they operate, where those devices are, what firmware they run, how they communicate, and who owns them will respond much faster than organizations discovering those facts during an incident.

CameraSwarm Is a Warning About Scale

More than 14,530 compromised devices show how quickly a single operator can turn an IoT weakness into a large campaign.

Zimbra Is a Warning About Speed

The Zimbra incident demonstrates that defenders can have very little time between public disclosure and real-world exploitation. The period immediately following disclosure should be treated as a high-risk window.

The Next Major Campaign May Not Need a New Zero-Day

Attackers already possess enormous inventories of known vulnerabilities, exposed devices, leaked credentials, and outdated infrastructure. The next major campaign could simply combine these existing weaknesses more efficiently.

✅ Confirmed: CERT Polska reported active exploitation of CVE-2026-73570 in Zimbra Collaboration and said unauthenticated attackers can execute shell commands as the zimbra user under the affected configuration.

✅ Confirmed: Zimbra identifies version 10.1.20 as containing the fix for CVE-2026-73570, while vulnerability records list the issue as an unauthenticated remote-code-execution vulnerability with a CVSS score of 8.9.

✅ Confirmed: Operation CameraSwarm involved more than 14,530 Dahua devices, with researchers identifying multiple compromise methods and approximately 1,923 devices carrying a persistent backdoor account.

Prediction

(+1) Emergency Zimbra Patching Will Accelerate

Organizations running vulnerable Zimbra installations are likely to prioritize emergency upgrades as active exploitation becomes more widely known. Security teams will also increase log hunting for evidence of previous compromise.

(+1) IoT Security Will Move Higher on Enterprise Security Agendas

The scale of CameraSwarm is likely to push more organizations toward formal security controls for cameras, recorders, and other connected physical infrastructure.

(+1) Network Segmentation Will Become More Important

Organizations will increasingly isolate IoT equipment from sensitive corporate networks because compromised edge devices can provide attackers with unexpected paths into enterprise environments.

(+1) Attackers Will Continue Reusing Old Vulnerabilities

Rather than relying exclusively on expensive zero-days, threat actors are likely to continue targeting old vulnerabilities where patching remains inconsistent and automated exploitation can produce thousands of victims.

(-1) Forgotten IoT Devices Will Remain a Major Weakness

Organizations that lack accurate inventories and firmware-management programs will continue to expose cameras and other appliances long after vulnerabilities become publicly known.

(-1) Exploited Zimbra Servers May Face Secondary Attacks

Once attackers gain command execution on vulnerable servers, compromised systems could become staging points for additional malware, credential theft, persistence, or lateral movement if organizations patch without performing post-exploitation investigation.

(+1) The Main Defensive Advantage Will Be Speed

The organizations most likely to avoid serious damage will be those that can identify vulnerable assets quickly, patch rapidly, isolate suspicious systems, and investigate exploitation before attackers establish durable persistence.

▶️ Related Video (72% Match):

🕵️‍📝Let’s dive deep and fact‑check.

🎓 Live Courses & Certifications:

Join Undercode Academy for Verified Certifications

🚀 Request a Custom Project:

Secure, high-velocity infrastructure and disruptive technological engineering. Contact our engineering team for high-tier development and proprietary systems:
[email protected]
💎 Smart Architecture | 🛡️ Secure by Design | ⭐ Trusted by Thousands

References:

Reported By: x.com
Extra Source Hub (Possible Sources for article):
https://www.facebook.com
Wikipedia
OpenAi & Undercode AI

Image Source:

Unsplash
Undercode AI DI v2

🔐JOIN OUR CYBER WORLD [ CVE News • HackMonitor • UndercodeNews ]

💬 Whatsapp | 💬 Telegram

📢 Follow UndercodeNews & Stay Tuned:

𝕏 formerly Twitter 🐦 | @ Threads | 🔗 Linkedin | 🦋BlueSky | 🐘Mastodon | 📺Youtube