Containing a Worm That Spreads Through USB Storage Devices

USB flash drives, external SSDs and memory cards can still bridge systems that were deliberately kept apart. A USB worm copies itself to removable media, then seeks another host when the device is connected. The result is not simply one infected laptop: shared drives and authorised transfer media can carry the malware between offices, suppliers and operational environments. In Australia, that risk is relevant wherever field data, production files or specialist software must move between a controlled network and a technician’s workstation.

Containing it requires more than deleting an alert or unplugging a drive. Teams must identify every potentially affected device and account, stop further copying, eradicate persistence, and prove that normal operations can resume without reviving the threat. CARM Security, developed by Exclusive Networks, fits this stage of the response by integrating signals from multiple security vendors so organisations can coordinate investigation, containment and post-breach remediation.

Why removable media remain a serious attack path

Removable-media worms have several common routes. One writes an executable and a launcher that appears to open the drive’s contents. Another modifies shortcuts so that opening a folder starts malicious code. A third exploits a weakness in the operating system or installed application, allowing the payload to execute with the user’s privileges. Some campaigns also copy data or add network-share routines, turning a removable-media infection into a broader intrusion. The USB device is the transport, but the damage is usually delivered through execution on the connected computer.

Autorun is no longer a dependable explanation for every incident. Current Windows releases restrict automatic execution from removable drives, yet users can still double-click a disguised shortcut, run an unexpected installer, approve a prompt or open a document that triggers a malicious action. Worm code may also alter the system so it runs after a reboot. That distinction matters during containment: teams should investigate execution, persistence and media modification rather than assume that disabled Autorun made USB attacks impossible.

The business case for restricted media is especially clear in Australian engineering, logistics and resource operations. A mining contractor in Kalgoorlie may carry a validated dataset on removable media to an isolated site network, while a defence supplier may move test files between facilities in Adelaide and Melbourne. These are legitimate transfers, not careless behaviour, and a blanket ban can delay critical work. Controls must distinguish approved, inspected devices from unknown media and track who used each one.

Detect the outbreak and establish its true scope

Start with evidence that survives the initial scramble. Endpoint detection and response records, antivirus alerts, USB device identifiers, file hashes, process activity and network logs can reveal copies of the worm and its execution chain. Ask whether the same storage device appears at more than one site, and whether identical files or registry changes appeared on unrelated systems. A single detected file hash is a lead, not a complete inventory: testing determines whether the infection moved before the first alert.

The first infection may have arrived through email rather than a drive already present on the network. A malicious message can deliver a lure or an initial payload. SPF alone is insufficient for blocking phishing emails, so email authentication should sit alongside DKIM, DMARC and phishing-resistant multifactor authentication rather than replace endpoint protection.

For Australian organisations, established national guidance can sharpen the response. The Australian Signals Directorate’s Essential Eight provides a useful baseline for application control, patching, user privileges, logging and network segregation. Critical infrastructure operators in energy, water, transport and telecommunications should also map removable-media pathways into operational technology, since a worm reaching an engineering workstation may attempt to cross into a control environment. Security teams should preserve relevant logs and request assistance through the Australian Cyber Security Centre when a broad or nationally significant incident is suspected.

Physical location matters too. Devices can arrive with staff or visitors through airports such as Sydney, Melbourne, Brisbane and Perth, while imported hardware moves through Australian Customs and Border Force processes. Border controls are not a substitute for endpoint vetting: an unknown drive arriving from overseas must enter the organisation through the same approval and scanning process as any other untrusted peripheral. Assign it an owner, quarantine it and test it in a controlled environment before approval.

Contain the worm while limiting operational disruption

Once the organisation can see the pattern, it can contain it with proportionate measures. Disconnect affected hosts from the network, including management access, and prevent them from reaching shared drives or operational systems. Do not move an infected drive between USB ports or machines, and do not plug it into a personal computer to test it. If forensic preservation is needed, keep the device powered and untouched only while the incident team directs; otherwise, an authorised shutdown limits continued execution.

Use device-control policy to block unauthorised mass-storage classes while preserving explicitly approved exceptions. Group Policy or endpoint management can limit which users may write to removable media, and the same control can prevent infected media from being reinserted after a reboot. Physical USB blockers can help in highly sensitive areas, but they are only one layer. A business exception should name the device, owner, permitted sites and expiry date, preventing a temporary operational need from becoming a permanent pathway for malware.

Network segmentation is particularly valuable when the affected machine supports site operations. Isolate engineering workstations, test benches and removable-media handling stations from production and corporate networks. Coordinate the action with system owners so a necessary update is not interrupted without a safe alternative. For organisations using CARM Security, a multi-vendor view of endpoint, identity and network alerts can help responders identify which hosts and accounts require immediate isolation instead of treating the entire estate as equally exposed.

Remove the infection and recover safely

Eradication begins with determining what the worm changed, copied or stole. A clean, trusted build is often safer and faster than removing every trace from a compromised system. Where an investigation depends on volatile memory or detailed adversary activity, preserve a forensic image before rebuilding. Removable media should be cleaned using a validated process or replaced; if its integrity cannot be established, retire it. Enterprise password changes, revoked tokens and session invalidation may be necessary if the malware enabled credential theft.

Recovery must use data and software that can be trusted. Reconcile restored files against known-good backups, scan them before they return to service, and avoid restoring executable files merely because they carry an expected filename. Re-enable authorised transfer routes only after representative systems have been tested. Endpoint detection content, detection rules and application allow-lists then need updating so that the removed strain, its artefacts and known persistence mechanisms are blocked across the environment.

The response also has legal and commercial consequences. Under the Australian Privacy Act and the OAIC’s Notifiable Data Breaches scheme, an incident involving personal information must be assessed for notification obligations, not assumed to be harmless because no ransom was demanded. Records should show what systems were affected, what data was accessed and what containment decisions were taken. Customers, regulators, cyber-insurance providers and contractual partners may all require different evidence, so a single factual timeline is more useful than disconnected updates.

Make the response repeatable and prove it works

Repeatable containment depends on knowing which media exist, who owns them and how they may move. Maintain an inventory of approved flash drives, external SSDs and devices used by contractors. Issue centrally managed, encrypted media where practical, require malware scanning on a dedicated inspection station, and provide a safe transfer process for engineers and suppliers. Staff training should cover unknown drives, shortcut tricks and the need to report a suspected worm immediately, rather than attempting to clean it themselves.

Test the process before an incident exposes its gaps. A tabletop exercise can follow a worm moving from a supplier’s drive to an Australian office, then across a segmented engineering environment, forcing participants to decide who isolates systems, who authorises shutdowns and who communicates impact. Include the security operations team, infrastructure owners, legal advisers, communications staff, business continuity managers and operational engineers. CARM Security can help connect multi-vendor alerts and response workflows, while the exercise tests whether people and authority—not just technology—can execute the plan.

Make the result an owned control rather than a one-off cleanup. Record the authoritative device list, exceptions and response contacts, then review alert coverage and recovery times after the exercise. The next step is to schedule a two-hour removable-media containment exercise for next week, using a benign media simulation and an authorised engineering workstation, with the security operations centre, infrastructure owner and communications lead assigned specific roles before the session begins.