Amazon identified a threat actor linked to the Democratic People’s Republic of Korea (DPRK) as the group behind multiple recent compromises of widely used open-source software libraries. The company warned that the attacks highlight a growing threat to software supply chains that underpin government and commercial IT systems.
Amazon Threat Intelligence said on July 29 that the DPRK-linked threat actor targeted popular Node Package Manager (NPM) libraries, including the axios, debug, chalk, and typo-crypto libraries.
This is the first time these various compromises are being publicly linked to one threat actor out of North Korea, CJ Moses, chief information security officer for Amazon Integrated Security, told reporters on Wednesday.
“For the DPRK, this is a financially motivated pattern,” Moses told reporters. “So, compromising a small number of highly popular packages gains potential access to thousands of downstream environments almost simultaneously.”
In March 2025, the DPRK-linked threat actor compromised the typo-crypto package. In September 2025, the same actor compromised the debug and chalk NPM packages.
In March 2026, the same operational playbook appeared in a compromise of the axios package, which is one of the most widely used JavaScript libraries. It is downloaded more than 100 million times each week.
“While the axios compromise has been publicly attributed to this group, the other three incidents have not been connected to that same threat actor, and that’s what we’re disclosing today is that we’ve tied them together,” Moses explained.
Moses said that the threat actor is using the same tactics across these different compromises.
In a July 29 company blog post, Moses said that in each case, the threat actor gained access by socially engineering a trusted maintainer of the package and then publishing a software update containing malicious code. Organizations that automatically downloaded the latest package versions received the compromised code.
Upon discovering this campaign, Amazon Threat Intelligence worked with Amazon Inspector to track, mitigate, and share the malicious package with the community through the Open Source Vulnerabilities database.
Amazon researchers also warned that generative artificial intelligence (AI) is helping adversaries to generate novel code and content at scale, making it more difficult for defenders to identify malicious packages.
“There are a lot more doors and windows for those adversaries to rattle and potentially find those vulnerabilities with their heightened vulnerability-finding capabilities,” said Rush Carskadden, director of application security at Amazon. “And the attack surface is exploding. The math just does not work out from an application security standpoint.”
“We have to give our security teams AI-driven capabilities to allow them to tilt that math back in their favor,” Carskadden said.