Cybersecurity researchers have flagged a significant supply chain attack targeting over a dozen packages associated with the GlueStack UI framework. Malicious code, which was introduced via a change to a core JavaScript file, allows attackers to run shell commands on infected machines. This malware also lets them take screenshots and even upload various files, giving them extensive remote access to any compromised systems. These affected packages collectively account for nearly one million weekly downloads, indicating a very widespread potential impact for many software developers. The first compromise in this particular campaign was reportedly detected by security firm Aikido Security on June 6, 2025, at night. This highlights the ongoing, active threats within the open-source software supply chain.
The malicious code injected into these packages is quite similar to a remote access trojan delivered in another recent npm compromise. This strong similarity suggests that the same persistent threat actors could be behind both of these widespread software supply chain attack campaigns. This particular trojan is an updated version that now supports new commands to harvest detailed system information and the public IP address. The project maintainers have since revoked the compromised access token and have marked all the impacted versions of packages as deprecated. However, attackers could still maintain access to machines that were infected before the malicious packages were eventually updated or fully removed. This creates a persistent risk for affected users.
This recent development comes as the security firm Socket discovered two separate rogue npm packages that masquerade as entirely legitimate utilities.
These packages, express-api-sync and system-health-sync-api, actually implant destructive wipers that can delete entire application directories on a system. One package, which is triggered by a secret key, executes a command to recursively delete all files from the current directory and below. The other, much more sophisticated package acts as both an information stealer and a wiper, using email as a covert communication channel.
This type of purely destructive attack is unusual as it appears motivated by sabotage rather than any direct financial gain from the attackers.
Concurrently, security researchers also discovered a new Python-based credential harvester, named imad213, on the Python Package Index (PyPI) repository. This malicious package, which has been downloaded thousands of times, deceptively claims to be a legitimate Instagram growth tool for many users. When it is executed by a victim, it prompts for their Instagram credentials and then immediately broadcasts them to ten different third-party bot services. The same threat actor has also uploaded other malicious packages to the PyPI repository that are designed to harvest credentials. These other packages target Facebook, Gmail, and Twitter accounts, or they can even be used to launch large-scale DDoS attacks against various platforms.
Reference:
