Amazon's security researchers have connected the compromises of several critical npm libraries, including **typo-crypto**, **debug**, **chalk**, and **axios**, to the notorious North Korean hacking group. This attribution, made with medium confidence, is based on shared tactics, techniques, and procedures (TTPs), command-and-control (C2) infrastructure, and operational similarities observed across the campaigns. ### A Timeline of Compromise The initial activity began in March 2025 with the trojanization of the **typo-crypto** package, which Amazon believes served as a testing ground for the attackers. The operation escalated significantly in September of the same year with the compromise of the widely used **debug** and **chalk** packages. This incident reportedly affected an estimated 10% of cloud environments within just two hours, highlighting the rapid propagation potential of such supply chain attacks. In March 2026, the attackers targeted **axios**, one of npm's most popular packages boasting over 100 million weekly downloads. While the **axios** incident had previously been publicly attributed to DPRK-linked actors, Amazon's analysis now connects it to the earlier package compromises, painting a clearer picture of **Sapphire Sleet**'s sustained campaign. ### The Modus Operandi **Amazon** states that the attackers gained access by socially engineering package maintainers. Once access was secured, malicious updates were published and automatically distributed to unsuspecting users. The primary motivation behind these attacks is believed to be financial, with the group targeting popular packages to gain indirect access to a large pool of potential downstream victims. ### Evolving Threat Landscape: Key Trends **Amazon** also highlighted several critical trends emerging from these recent supply chain attacks, which pose significant challenges for IT security professionals and privacy-conscious users: * **Distributed Malicious Functionality**: Attackers are splitting malicious code across multiple seemingly benign packages, making detection more difficult. * **Trust Building**: Threat actors are spending months building trust by maintaining legitimate projects or becoming contributors before introducing malicious code. * **Decoupled Malicious Behavior**: Malicious behavior is increasingly decoupled from package contents, relying on external scripts, configuration files, or servers that can be weaponized later. * **Advanced Malware Techniques**: Malware is using stronger encryption and multi-stage payloads, often with runtime or remotely fetched keys that hinder static analysis. * **Environment Awareness**: Payloads are becoming environment-aware, delaying execution unless they detect real developer or production environments to evade analysis sandboxes. * **Slopsquatting**: Attackers are increasingly exploiting "slopsquatting" by registering package names hallucinated by AI coding assistants, hoping developers or autonomous coding agents will install them. **Amazon** further explains that many of these tactics are enhanced and simplified by AI, as it helps attackers generate code, documentation, and maintainer identities with greater ease. ### Industry Response In response to these growing dangers, **Amazon** has outlined a multi-faceted approach. This includes reporting its findings and intelligence to the community, collaborating with organizations like **OpenSSF** and other industry partners, and investing $12.5 million in the **Akrites** initiative, which aims to protect critical open-source software from AI-enabled attacks. These efforts underscore the collaborative and proactive measures required to combat the evolving threat landscape in open-source software.