QEMU is an open-source platform that provides a secure and private virtualized space for trying out malicious codes, exploits, and attacks on their own environments.
This controlled testing ground minimizes the risk of detection and legal matters.
Moreover, QEMU permits hackers to develop malware that can run across different hardware architectures and operating systems.
Cybersecurity researchers at Kaspersky Labs recently discovered that hackers are abusing the QEMU hardware emulator to covertly exfiltrate stolen data.
Technical analysis
Attackers love using genuine tools in order to avoid detection, but also to reduce malware expenditure.
Data exfiltration, drive encryption, remote execution, and memory dumping are some of the network scanning activities that trusted software supports.
Pre-installed malware or employee-mimicking RDP/VPN access act as footholds through compromised systems.
Network tunnels and port forwarding utilities enable users to bypass NAT and firewalls thus gaining entry into the internal system.
Numerous tools exist for creating network tunnels between systems, some direct, others using proxies to mask attacker IPs. These utilities frequently surfaced during researchers’ incident response efforts over the past three years.
Here below we have mentioned the tools:-
- Stowaway
- ligolo
- 3proxy
- dog-tunnel
- chisel
- FRP
- ngrok
- gs-netcat
- plink
- iox
- nps
During a recent investigation, experts uncovered suspicious activity in a company’s system. Among the tools detected were Angry IP Scanner, mimikatz, and QEMU.
The presence of QEMU puzzled security analysts – why would attackers use a virtualizer? Further analysis revealed an unusual QEMU execution command line without a LiveCD or disk image.
Here below we have mentioned all the arguments:-
- m 1M: Allocated only 1MB RAM, insufficient for most OSes.
- netdev user,id=lan,restrict=off: Created virtual network interface ‘lan’ to communicate externally, without restrictions.
- netdev socket,id=sock,connect=<IP>:443: Socket interface ‘sock’ connected to remote server at <IP>:443.
- netdev hubport,id=port-lan/port-sock,hubid=0,netdev=lan/sock: Added ports to virtual hub linked to ‘lan’ and ‘sock’ interfaces.
- nographic: Started QEMU in non-GUI console mode.
The external <IP> raised suspicions, and QEMU allows VM interconnections via -netdev options by creating network backends.
Network tunnel (Source – Securelist)
Experts used QEMU to build a Kali Linux VM on AttackerServer. The VM was connected through a socket adapter and listened on port 443. PivotHost had another QEMU instance which connected to the socket over AttackerServer’s port 443.
The user adapter was conjoined via the hub with the socket. The adversary’s QEMU options were imitated. A PivotHost-AttackerServer tunnel was set up by QEMU to enable subnet scans from PivotHost to the Kali VM.
QEMU does not encrypt tunneled traffic, as it sends unencrypted encapsulated packets. The packet data contains the Ethernet frame size and frame. Intercepted traffic can be obtained by removing headers.
However, besides this, threat actors use legitimate tools ingeniously. Multi-level protection with endpoint and network monitoring by SOC experts is crucial for timely rarity detection and attack blocking. While the MDR service can detect suspicious QEMU activity.