After detecting suspicious activity on its Networked Experimentation, Research, and Virtualization Environment (NERVE), a collaborative network used for research, development, and prototyping, security experts discovered a compromise by a foreign nation-state threat actor.
MITRE is a federally funded research and development organization that supports various U.S. government agencies in the aviation, defense, healthcare, homeland security, and cybersecurity fields.
The attack on MITRE came through the exploitation of two Ivanti Connect Secure Zero Day vulnerabilities tracked as CVE-2023–46805 and CVE-2024–21887, respectively.
NERVE Offline
Following detection of the incident, MITRE took action to contain the incident, including taking the NERVE environment offline, and quickly launched an investigation with the support of in-house and leading third-party experts.
Additionally, on December 31, the adversary deployed the ROOTROT web shell on an external-facing Ivanti appliance, gaining initial access to NERVE, according to a blog post on Medium.com. This early intrusion leveraged the Ivanti Zero Days for unauthorized access before the initial disclosure of vulnerabilities on January 10 and before patches were available. By leveraging this access point, the adversary infiltrated the NERVE network, circumventing multi-factor authentication, and established a foothold for subsequent activities.
The subsequent hijacking of sessions and utilization of RDP over HTML5 capabilities allowed the adversary to establish connections to systems within the NERVE.
On January 4, the attacker profiled MITRE’s NERVE environment, interacting with vCenter from the compromised Ivanti appliance, establishing communication with multiple ESXi hosts, according to the post. Subsequently, they successfully logged into several accounts within the NERVE via RDP, leveraging hijacked credentials to access user bookmarks and file shares to gain insights into the network architecture.
Furthermore, the adversary on Jan. 5 manipulated VMs and established control over the infrastructure, according to the report. The adversary used compromised administrative credentials, authenticated from an internal NERVE IP address, indicating lateral movement within the NERVE. Additionally, they attempted to enable SSH and attempted to destroy one of their own VMs as well as POSTed to /ui/list/export and downloaded a file demonstrating a sophisticated attempt to conceal their presence and maintain persistence within the network.
Malicious Payloads
Meanwhile, the adversary on Jan. 7 accessed VMs and deployed malicious payloads, including the BRICKSTORM backdoor and a web shell MITRE called BEEFLUSH. These actions established persistent access and allowed the adversary to execute arbitrary commands and communicate with command-and-control servers. The adversary utilized techniques such as SSH manipulation and execution of suspicious scripts to maintain control over the compromised systems.
A VMware default account vpxuser, used VMware vSphere Management API pyvmomi, made seven API calls that enumerated a list of mounted and unmounted drives. The adversary pivoted back to the admin account and created three new VMs, all conforming to the local naming convention, and successfully logged into them from an internal IP address. One of these VMs was deleted on the same day.
BRICKSTORM was in VMs in the /mnt/cpt directory named tmpd and in the /bin directory named httpd (/mnt/cpt/tmpd and/bin/httpd). Both versions were given local persistence mechanisms. /mnt/cpt/tmpdwas given both /etc/rc.local and/etc/init.d/urandom_seed while/bin/httpd was given /etc/init.d/urandom_seed persistence method. BRICKSTORM communicated with the C2 domains seen in Appendix 1.
BEEFLUSH, /idm/..;/resources/css/defaultb.jsp, communicated with several internal IP addresses making POST requests. While in the vCenter server, the adversary executed suspicious Python scripts and /bin/sh commands from the /tmp directory.
On Jan. 10, the zero-day vulnerabilities were publicly disclosed via Ivanti Advisory.
Additionally, on Jan. 11, according to an analysis of system memory the adversary used the Ivanti help website as a staging area for data exfiltration, the post said. They made requests to /dana-na/help/ on the Ivanti appliance, where a base64 encoded logo.gif file was an exact copy of a log file on the system which the adversary exfiltrated.
The adversary uploaded a Python script, visits.py, that contained the WIREFIRE (aka GIFTEDVISITOR) web shell on the Ivanti appliance within the /home/venv3/lib/Python3.6/site-packages/cav-0.1-py3.6.egg archive file.
On Jan. 12, new advisories published from CISA and Mandiant.
Data Exfiltrated
On Jan. 19, the adversary exfiltrated data from the NERVE using command-and-control infrastructure, according to the post. An external IP address, 172.75.64[.]253, made network traffic requests to the BUSHWALK web shell, /dana-na/jam/querymanifest.cgi.
From February to mid-March, the adversary attempted lateral movement and maintained persistence within the NERVE. Despite unsuccessful attempts to pivot to other resources, the adversary persisted in accessing other virtual environments within vCenter.
The adversary executed a ping command for one of MITRE’s corporate domain controllers and attempted to move laterally into MITRE systems but was unsuccessful.
The investigation into the attack is ongoing.
“No organization is immune from this type of cyberattack, not even one that strives to maintain the highest cybersecurity possible,” said Jason Providakes, president and chief executive of MITRE after release of the attack in April. “We are disclosing this incident in a timely manner because of our commitment to operate in the public interest and to advocate for best practices that enhance enterprise security as well necessary measures to improve the industry’s current cyber defense posture. The threats and cyberattacks are becoming more sophisticated and require increased vigilance and defense approaches. As we have previously, we will share our learnings from this experience to help others and evolve our own practices.”

