Recent advancements in tools have enhanced the detection of common exploitation tactics for open-source packages, yet one critical aspect remains underexamined: entry points.

Security experts from Checkmarx have revealed that attackers can exploit entry points across multiple programming ecosystems, particularly focusing on PyPI, to deceive users into executing malicious code. Although this technique does not lead to immediate system compromise, it represents a stealthier strategy for attackers to infiltrate systems, potentially bypassing conventional security measures.

Entry points are designed to expose package functionalities, but they also present vulnerabilities across several ecosystems, including PyPI (Python), npm (JavaScript), Ruby Gems, NuGet (.NET), Dart Pub, and Rust Crates. Malicious actors can exploit these entry points to run harmful code when specific commands are executed, creating a widespread threat.

The researchers identified various attack vectors, including “command-jacking,” which involves impersonating popular third-party tools and system commands, as well as targeting different stages of the development cycle with harmful plugins and extensions. Each method varies in its likelihood of success and potential for detection.

https://www.youtube.com/watch?v=-GTOOrYniFE

One particularly deceptive approach is known as “command wrapping.” This technique does not simply replace a command; instead, it creates an entry point that serves as a wrapper around the legitimate command. When the user executes the command, the malicious entry point activates and runs the attacker’s code before passing the original command along with any user-provided arguments. This method is especially concerning as it gives the illusion of normal operation, making detection through standard usage exceedingly difficult.

https://www.youtube.com/watch?v=UclRQxQ0Wbg&ab_channel=CheckmarxZero

Additionally, widely used development tools, such as Flake8, could also be vulnerable. An attacker might create a harmful extension disguised as useful linting rules, allowing them to perform detrimental actions on the victim’s machine, inject malicious “fixes” into the code, or manipulate linting outcomes to obscure or fabricate issues.

The researchers also pointed out that the rising use of Python wheels (.whl files) introduces a unique challenge. While these files do not execute setup.py during installation, complicating arbitrary code execution for attackers, the entry point attack strategy provides a means to circumvent this limitation.

“Many security tools primarily focus on analyzing the execution of preinstall scripts during installation, which are usually associated with .tar.gz files,” the researchers noted. “Consequently, they might overlook malicious code in packages distributed as .whl files, especially when the harmful behavior is activated through entry points instead of direct execution.”

The researchers stressed the necessity for comprehensive security measures that address the risks associated with entry points. By understanding and mitigating these vulnerabilities, the industry can work toward a more secure Python packaging ecosystem, protecting both individual developers and enterprise systems from sophisticated supply chain attacks.