Software supply chain attacks are escalating at an alarming rate

By Ross Bryant

[By Ross Bryant, Chief of Research at Phylum]

If there is one safe prediction that I can make in 2024, it is that software supply chain attacks will continue to grow at an alarming rate. My team’s job is to track bad actors across the open-source software ecosystem, and there was a lot to see in 2023. Our Q4 2023 research report revealed that the software supply chain is one of the easier and more popular attack vectors. This vector is an easy target since open source is used in 97% of projects and included in more than 70% of code bases.  The research discovered a significant increase in targeted organizations and attack sophistication, especially within financial and cryptocurrency organizations, with monetary gain as a top motivator. 

As 2024 evolves, popular attack methods such as production system credential theft and financial resources (e.g., personally identifiable information (PII), cryptocurrency, etc.) will remain top threats. Attackers will also continue to execute ransomware-style campaigns – leveraging access to customer data and assets and using the threat of stolen information to coerce organizations to pay the ransom. 

A Surprise in the Numbers

This quarterly research report showed a slight decrease in published packages compared to the previous quarter. However, the number of targeted organizations increased substantially—262.63% more targeted attacks compared to Q3 2023, which had risen 47% from Q2 2023. This showed a clear trend across 2023 of increased direct, targeted attacks.

While the number of published packages was lower than in previous reports, a larger portion focused on specific organizations and indicated specific methods associated with software supply chain threats. 

Mistaken identity 

One such attack method, dependency confusion, is a software supply chain attack that exploits a state of confusion in package managers – checking for named packages within a public registry first before searching in a private registry. An attacker can register an identically named, malicious package on the public registry, intending for the package manager to inadvertently download it, mistaking it for the legitimate package.

Another method that exploits public registries is brandsquatting. In this method, threat actors use popular brand names to mask their malicious code and lure, mislead, and trick the developer into downloading the malicious package.

Attacker Subtly = Greater Gains

In this recent research, two common approaches emerged for targeting the software supply chain: production system credential theft and stealing financial resources (e.g., bank account information, cryptocurrency, etc.). 

In one attack, a threat actor targeted a select group of widely used cloud provider SDKs. A review of the code revealed that the attacker was specifically interested in sensitive credentials to cloud infrastructure.

Exploiting developers’ trust in these packages, the attacker slightly modified a vital part of the code responsible for managing and handling credentials. This triggered a stealthy HTTP POST request for the users’ access and secret keys to a remote URL under the attacker’s control. By making subtle changes and republishing these altered packages on PyPI with similar names, the attacker blended in to remain undetected while maintaining the packages’ expected functionality.

This method, used in at least five packages, involved a simple and effective technique to obscure the remote URL, demonstrating a calculated approach to infiltrating trusted software components on developer workstations and production infrastructure.

Some Organizations Take Proactive Security Steps 

In Dec 2023, an article was published outlining the discovery of an additional set of oddly sophisticated packages. Unlike some of the other campaigns, this one was highly targeted.

These packages contained an encrypted component that could only be unlocked with data from the environment of a local machine in a specific network, where the decryption key was the hostname of a particular organization. Once decrypted, the payload was executed, and user credentials were moved laterally inside the network to a Microsoft Teams Webhook. This left few options: a threat actor had gained a deep foothold in the network, this was a security audit, or this was the work of an insider threat.

Realizing these packages’ specific focus, the targeted organization was contacted to warn and mitigate an attack. If this were an external threat actor, the organization needed to be notified of it before the attacker could do considerable damage.

The analysis continued to explore a very advanced and sophisticated attack comparable with other APT (Advanced Persistent Threat) campaigns.

However, once contact was established with the targeted company, it was discovered that this was part of a broad internal security assessment aimed at mimicking pressing real-world threats. The mimicked attack looked to replicate behaviors the organization was seeing from attackers leveraging the software supply chain as a conduit into their network.

Why Organizations Should Prioritize Software Supply Chain Security

In 2024, attackers will become even more sophisticated, finding new ways to access an organization’s valuable customer and corporate data by exploiting the software supply chain. Methods such as dependency confusion and brandsquatting are the beginning, easily fooling package managers and developers alike. 

Heightened focus on the software supply chain should be a critical component of an organization’s security portfolio, especially those in the financial and cryptocurrency arenas.


No posts to display