Cybersecurity researchers have identified a previously undocumented malware strain called KadNap that is primarily infecting Asus routers and other internet-facing networking devices. The attackers are using these compromised systems to form a botnet that routes malicious traffic through residential connections, effectively turning infected hardware into anonymous proxy nodes.
The threat was first observed in real-world attacks in August 2025. Since that time, the number of affected devices has grown to more than 14,000, according to investigators at Black Lotus Labs. A large share of infections, exceeding 60 percent, has been detected within the United States. Smaller groups of compromised devices have also been identified across Taiwan, Hong Kong, Russia, the United Kingdom, Australia, Brazil, France, Italy, and Spain.
Researchers report that the malware uses a modified version of the Kademlia Distributed Hash Table (DHT) protocol. This peer-to-peer networking technology enables the attackers to conceal the true location of their infrastructure by distributing communication across multiple nodes. By embedding command traffic inside decentralized peer-to-peer activity, the operators can evade traditional network monitoring systems that rely on detecting centralized servers.
Within this architecture, infected devices communicate with one another using the DHT network to discover and establish connections with command-and-control servers. This design improves the botnet’s resilience, as it reduces the chances that defenders can disable operations by shutting down a single control point.
Once a router or other edge device has been compromised, the system can be sold or rented through a proxy platform known as Doppelgänger. Investigators believe this service is a rebranded version of another proxy operation called Faceless, which previously had links to TheMoon router malware. According to information published on the Doppelgänger website, the service launched around May or June 2025 and advertises access to residential proxy connections in more than 50 countries, promoting what it claims is complete anonymity for users.
Although many of the observed infections involve Asus routers, researchers found that the malware operators are also capable of targeting a wider range of edge networking equipment.
The attack chain begins with the download of a shell script named aic.sh, retrieved from a command server located at 212.104.141[.]140. This script initiates the infection process by connecting the compromised device to the botnet’s peer-to-peer network.
To ensure the malware remains active, the script establishes persistence by creating a cron task that downloads the same script again at the 55-minute mark of every hour. During this process, the file is renamed “.asusrouter” and executed automatically.
After persistence is secured, the script downloads an ELF executable, renames it “kad,” and runs it on the device. This program installs the KadNap malware itself. The malware is capable of operating on hardware that uses ARM and MIPS processor architectures, which are commonly found in routers and networking appliances.
KadNap also contacts a Network Time Protocol (NTP) server to retrieve the current system time and store it along with the device’s uptime. These values are combined to produce a hash that allows the malware to identify and connect with other peers within the decentralized network, enabling it to receive commands or download additional components.
Two additional files used during the infection process, fwr.sh and /tmp/.sose, contain instructions that close port 22, which is the default port used by Secure Shell (SSH). These files also extract lists of command server addresses in IP-address-and-port format, which the malware uses to establish communication with control infrastructure.
According to researchers, the use of the DHT protocol provides the botnet with durable communication channels that are difficult to shut down because its traffic blends with legitimate peer-to-peer network activity.
Further examination revealed that not every infected device communicates with every command server. This suggests the attackers are segmenting their infrastructure, possibly grouping devices based on hardware type or model.
Investigators also noted that routers infected with KadNap may sometimes contain multiple malware infections simultaneously. Because of this overlap, it can be challenging to determine which threat actor is responsible for particular malicious activity originating from those systems.
Security experts recommend that individuals and organizations operating small-office or home-office (SOHO) routers take several precautions. These include installing firmware updates, restarting devices periodically, replacing default administrator credentials, restricting management access, and replacing routers that have reached end-of-life status and no longer receive security patches.
Researchers concluded that KadNap’s reliance on a peer-to-peer command structure distinguishes it from many other proxy-based botnets designed to provide anonymity services. The decentralized approach allows operators to remain hidden while making it significantly harder for defenders to detect and block the network.
In a separate report, security analysts at Cyble disclosed a new Linux malware threat named ClipXDaemon.
The malware targets cryptocurrency users by intercepting wallet addresses that victims copy to their clipboard and secretly replacing them with addresses controlled by attackers. This type of threat is commonly known as clipper malware.
ClipXDaemon is distributed through a Linux post-exploitation framework called ShadowHS and has been described as an automated clipboard-hijacking tool designed specifically for systems running Linux X11 graphical environments.
The malware operates entirely in memory, which reduces traces on disk and improves its ability to remain undetected. It also employs several stealth techniques, including disguising its process names and deliberately avoiding execution in Wayland sessions.
This design choice is intentional because Wayland’s security architecture introduces stricter restrictions on clipboard access. Applications must usually involve explicit user interaction before they can read clipboard contents. By disabling itself when Wayland is detected, the malware avoids triggering errors or suspicious behavior.
Once active in an X11 session, ClipXDaemon continuously checks the system clipboard every 200 milliseconds. If it detects a copied cryptocurrency wallet address, it immediately substitutes it with an attacker-controlled address before the victim pastes the information.
The malware currently targets a wide range of digital currencies, including Bitcoin, Ethereum, Litecoin, Monero, Tron, Dogecoin, Ripple, and TON.
Researchers noted that ClipXDaemon differs significantly from traditional Linux malware families. It does not include command-and-control communication, does not send beaconing signals to remote servers, and does not rely on external instructions to operate.
Instead, the malware generates profits directly by manipulating cryptocurrency transactions in real time, silently redirecting funds when victims paste compromised wallet addresses during transfers.