top of page
perceptive_background_267k.jpg

18 Malicious npm Packages Deliver Cross-Platform RAT to Alibaba Tool Users

Published:

3 August 2026 at 18:43:53

Alert date:

3 August 2026 at 19:01:35

Source:

thehackernews.com

Click to open the original link from this advisory

Supply Chain & Dependencies, Ransomware & Malware, Data Breach & Exfiltration

Cybersecurity researchers discovered 18 malicious npm packages targeting users of Alibaba developer tools with a cross-platform remote access trojan (RAT). The attack is described as a sophisticated, targeted software supply chain attack aimed at Chinese-speaking development environments. One of the identified packages, 'lib-mtop', is an unscoped package that mimics the name of a legitimate private Alibaba package, indicating a typosquatting or namespace confusion strategy. The campaign appears specifically crafted to blend into the Alibaba developer ecosystem to maximize infection rates. The RAT deployed is cross-platform, suggesting it can infect Windows, macOS, and Linux systems. This attack highlights the growing risk of supply chain attacks via public package registries like npm. The targeting of Chinese-speaking environments and Alibaba-specific tooling suggests a focused threat actor with knowledge of the target ecosystem.

Technical details

Cybersecurity researchers discovered 18 malicious npm packages targeting users of Alibaba developer tools with a cross-platform RAT. The campaign uses a multi-layer dependency tree attack: top-layer lure packages impersonate private @ali-scoped Alibaba packages, a middle-layer bridge package ('smart-config-manager') connects them to low-layer packages containing loader logic. The loader fetches a remote JavaScript payload via curl. A low-layer package contacts a GitHub repository to retrieve a rule engine configuration, which then uses Node.js's vm module to execute a malicious payload that contacts a remote server for secondary malware. The payload is retrieved from a domain masquerading as Alibaba ('aone-cli-next.oss-cn-beijing.aliyuncs[.]com'). Platform-specific behaviors: On Windows, it terminates Alilang (enterprise security/VPN/office app) and replaces its core code with a trojanized version; On Linux, it downloads a binary to /tmp, runs it as a detached process, then deletes it from disk after loading into memory; On macOS, it inserts a malicious script into ~/.zshrc and sets up a 10-minute Launch Agent. The final payload is a complex backdoor with command execution, arbitrary file upload/download, host reconnaissance, payload staging, and lateral movement capabilities. It persists by injecting malicious code into enterprise collaboration apps like DingTalk, Wukong, and Qoder. Chinese language comments in source code and GitHub commit timestamps at UTC+08:00 suggest a Chinese-speaking threat actor. The campaign goal appears to be industrial espionage. A secondary incident involves a poisoned PyPI package 'mrmustard' (v0.7.4) that steals SSH private keys, AWS credentials, Kubernetes configurations, SLURM job queues, and GPU inventories, exfiltrating them to 'metrics.femboy[.]energy'. The mrmustard compromise was achieved by breaching the maintainer's GitHub account via self-hosted CI runners and exfiltrating publishing secrets to a webhook.site URL. The malware installed three separate persistence mechanisms that survive package uninstallation.

Mitigation steps:

1. Immediately uninstall any of the 18 identified malicious npm packages if present in your environment. 2. Assume compromise if any of the listed packages were installed; rotate all sensitive credentials (API keys, tokens, passwords) from a clean, unaffected machine. 3. Audit developer systems for signs of suspicious activity, including unauthorized processes, modified application binaries (especially Alilang), changes to ~/.zshrc (macOS), and unexpected Launch Agents. 4. On Linux, check /tmp for suspicious binaries and review running detached processes. 5. Inspect DingTalk, Wukong, and Qoder application files for injected malicious code. 6. Block network connections to 'aone-cli-next.oss-cn-beijing.aliyuncs[.]com' and 'metrics.femboy[.]energy'. 7. For mrmustard PyPI users: uninstall version 0.7.4 immediately, rotate SSH private keys, AWS credentials, and Kubernetes configurations from a clean machine, and check for persistence mechanisms that survive package uninstallation. 8. Review CI/CD pipeline secrets and rotate any that may have been exposed, particularly for self-hosted runners. 9. Implement npm package integrity checks and prefer scoped (@ali) packages over unscoped equivalents when working with Alibaba tooling. 10. Monitor for outbound connections to suspicious domains masquerading as Alibaba infrastructure.

Affected products:

lib-mtop (npm package
versions v1.0.1
v1.0.2
v1.0.3)
aone-kit (npm package)
aone-kit-cli (npm package)
aone-sandbox (npm package)
local-config-parser (npm package)
smart-config-manager (npm package)
cloud-config-fetcher (npm package)
fast-transform-pipeline (npm package)
aone-cloud-cli (npm package)
colder-cli (npm package)
def-open-client (npm package)
feedback-ai-sdk (npm package)
flight-compare-analyzer (npm package)
lwp-web-client (npm package)
lzd-unified-station-sdk (npm package)
open-worker-cli (npm package)
test-skill-zip (npm package)
uniapi-bridge (npm package)
mrmustard (PyPI package
version 0.7.4)
Alilang (enterprise security/VPN/office app
Windows)
DingTalk (enterprise collaboration app)
Wukong (enterprise collaboration app)
Qoder (enterprise collaboration app)

Related links:

Related CVE's:

Related threat actors:

IOC's:

aone-cli-next.oss-cn-beijing.aliyuncs[.]com, metrics.femboy[.]energy, npm maintainer account: ch4ce, npm package: lib-mtop, npm package: aone-kit, npm package: aone-kit-cli, npm package: aone-sandbox, npm package: local-config-parser, npm package: smart-config-manager, npm package: cloud-config-fetcher, npm package: fast-transform-pipeline, npm package: aone-cloud-cli, npm package: colder-cli, npm package: def-open-client, npm package: feedback-ai-sdk, npm package: flight-compare-analyzer, npm package: lwp-web-client, npm package: lzd-unified-station-sdk, npm package: open-worker-cli, npm package: test-skill-zip, npm package: uniapi-bridge, PyPI package: mrmustard==0.7.4, webhook[.]site (used for exfiltrating publishing secrets in mrmustard incident), File path: ~/.zshrc (macOS persistence), Directory: /tmp (Linux payload drop)

This article was created with the assistance of AI technology by Perceptive.

© 2025 by Perceptive Security. All rights reserved.

email: info@perceptivesecurity.com

Disclaimer: Deze website toont informatie afkomstig van externe bronnen. Perceptive aanvaardt geen verantwoordelijkheid voor de inhoud, juistheid of volledigheid van deze informatie.

bottom of page