Cybersecurity researchers from SafeDep and Veracode have uncovered sophisticated malware campaigns targeting developers through compromised npm packages, revealing increasingly complex attack methods in the software supply chain.

## Four Malicious Packages Removed from npm

Four malicious npm packages were identified and removed from the registry, but not before accumulating significant downloads:

– **eslint-config-airbnb-compat** (676 downloads)
– **ts-runtime-compat-check** (1,588 downloads)
– **solders** (983 downloads)
– **@mediawave/lib** (386 downloads)

## Complex Multi-Stage Attack Chain

The **eslint-config-airbnb-compat** package employed a sophisticated transitive dependency attack. It listed ts-runtime-compat-check as a dependency, which then contacted an external server to retrieve and execute Base64-encoded malicious code.

The **solders** package used an even more elaborate approach, featuring obfuscated JavaScript that appeared as random Japanese Unicode characters. Once decoded, the malware:

1. Checks if the target system runs Windows
2. Executes PowerShell commands to download additional payloads
3. Configures Windows Defender exclusions to avoid detection
4. Downloads a .NET DLL that extracts hidden data from PNG image pixels
5. Ultimately deploys Pulsar RAT, a remote administration tool

This attack chain involved over a dozen layers of obfuscation, demonstrating extraordinary efforts to evade detection.

## Cryptocurrency-Focused Threats Surge

Socket researchers identified a growing trend of malware targeting cryptocurrency and blockchain developers, including:

– **Credential stealers** harvesting browser and wallet credentials
– **Cryptocurrency drainers** that empty victims’ wallets using multi-hop transfers
– **Clippers** that monitor clipboards and replace wallet addresses with attacker-controlled ones

## AI Coding Creates New Attack Vector

The rise of AI-assisted coding has introduced “slopsquatting” attacks, where large language models hallucinate non-existent package names that attackers can then register maliciously. While enhanced AI agents can reduce these phantom suggestions by approximately 50%, they cannot eliminate the risk entirely.

## Red Team Exercise Mistaken for Real Threat

A package called **chimera-sandbox-extensions** on PyPI was initially identified as sophisticated malware targeting corporate environments and macOS systems. The package appeared designed to steal:

– JAMF configuration data
– CI/CD environment variables
– AWS tokens and credentials
– General system information

However, Grab later clarified that this was actually a controlled red team exercise conducted by their cybersecurity team to test internal security capabilities, not a genuine threat.

## Key Takeaways

These incidents highlight the evolving sophistication of supply chain attacks targeting developers. Organizations should implement robust package verification processes, monitor dependencies carefully, and maintain updated security practices to defend against these increasingly complex threats.

The multi-layered obfuscation techniques and targeted approaches demonstrate that attackers are investing significant resources in bypassing security measures, making vigilance and proactive security research more critical than ever.