Jenkins has been discovered with a critical vulnerability that is associated with arbitrary code execution that threat actors can exploit for malicious purposes.
The vulnerability is tracked as CVE-2024-23898, and the severity is yet to be categorized.
However, Jenkins has fixed this vulnerability on their latest versions, 2.442 and LTS 2.426.3. Jenkins currently has a market share of 44% as of 2023, which means that the potential impact of exploiting this vulnerability can be catastrophic.
AI-Powered Protection for Business Email Security
Trustifi’s Advanced threat protection prevents the widest spectrum of sophisticated attacks before they reach a user’s mailbox. Try Trustifi Free Threat Scan with Sophisticated AI-Powered Email Protection .
CVE-2024-23898: Cross-Site WebSocket hijacking
This particular vulnerability exists due to the lack of origin validation of requests that are made through the CLI WebSocket endpoint, which could allow a threat actor to execute CLI commands on the Jenkins controller. Moreover, the vulnerability has specific criteria which must be met for an attacker to exploit it.
This vulnerability is dependent on the fact that most browsers do not implement a “lax by default” policy, which serves as a potential safeguard against this vulnerability. Nevertheless, the exploitation of this vulnerability requires a malicious link to be sent to the victim, and a user interaction is mandatory.
According to the reports shared with Cyber Security News by SonarSource, one of the methods to invoke the Jenkins-CLI command is by using web sockets.
Jenkins-CLI allows users to execute custom commands that are implemented in the Hudson/CLI directory of the Jenkins Git repository.
The most common way of invoking a command is using Jenkins-cli.jar or SSH. However, there was another method discovered that invokes commands by sending two POST requests to http://jenkins/cli?remoting=false.
When a CLI command is invoked, Jenkins uses args4j’s parseArgument, which calls expandAtFiles. If an attacker can control the arguments, it can be expanded to an arbitrary number of ones from an arbitrary file on the Jenkins instance.
Exploitation
Moreover, this arbitrary command execution combines a Data-leak vulnerability (CVE-2024-23897) with a similar background. CVE-2024-23897 exists due to a similar reason but is associated with leaking the contents of an arbitrary file on the Jenkins instance.
Numerous researchers have replicated several attack scenarios and have produced functional proof-of-concept (PoC) exploits that have been made available to the public on GitHub.
However, this Data-leak vulnerability is leveraged for reading the contents of the file in order to find the arbitrary number of arguments, which is later used in the exploitation of Remote Code Execution.
In addition, the success rate of arbitrary code execution over the Jenkins instance is dependent on different contexts. Some of the interesting files that are evident for successful exploitation are
- SSH keys
- /etc/passwd, /etc/shadow
- Project secrets and credentials (refer to Jenkins’ advisory for more information)
- Source code, build artifacts
- and more…
