Security researchers have uncovered a new cyber-threat targeting publicly exposed instances of the Docker Engine API. Attackers exploit misconfigurations to deploy a malicious Docker container housing Python malware compiled as an ELF executable, functioning as a Distributed Denial of Service (DDoS) bot agent. The Docker Engine API, previously targeted, has gained popularity for initiating such attacks, often linked with cryptojacking malware.
In this campaign, attackers initiate access with an HTTP POST request to Docker’s API, retrieving a malicious Docker container from Dockerhub, cleverly disguised as a seemingly innocuous MySQL image for Docker.
A static analysis of the malware’s ELF executable reveals a 64-bit, statically linked ELF with intact debug information, indicating Python code compiled with Cython. The code focuses on various DDoS methods, including SSL-based, UDP-based, and Slowloris-style attacks. The bot connects to a command-and-control (C2) server, authenticating with a hard-coded password, and Cado Security Labs monitored botnet activity, witnessing DDoS attacks using UDP- and SSL-based floods.
While no actual mining activity was observed, the researchers caution that the malicious container contains files that could facilitate such actions. Although not categorized as a supply chain attack, Docker Hub users are urged to remain vigilant, perform periodic assessments of pulled images, and implement network defenses.
Cado Security Labs reported the malicious user behind OracleIV to Docker, emphasizing the ongoing existence of malicious container images in Docker’s library. Users are encouraged to stay proactive in mitigating risks associated with misconfigured internet-facing services.
The novel campaign highlights the significance of securing Docker Engine APIs against potential threats arising from misconfigurations and reinforces the need for continuous monitoring and defensive measures to protect against evolving cyber threats.