npm Mass Typosquat Campaign Delivering Cross-Platform RAT and Infostealer

Technical Analysis

The campaign leverages large-scale npm package publishing with names resembling legitimate popular libraries, exploiting typosquatting and possibly AI-slop naming patterns to evade manual review and automated scanning heuristics. Once installed, the packages execute a payload that establishes RAT functionality alongside credential and data exfiltration (infostealer) capabilities across Windows, macOS, and Linux hosts. No specific CVE has been assigned; the attack vector is package installation via npm install rather than a software vulnerability exploit. Because npm packages are frequently pulled into CI/CD pipelines, developer environments, and increasingly into AI coding assistants and autonomous coding agents that auto-resolve and install dependencies, this campaign poses a direct risk to AI agent systems that perform unsupervised package installation, potentially exposing API keys, cloud credentials, and other secrets used by agent tooling. Organizations running LLM-based coding agents or RAG pipelines that ingest npm packages as tools/dependencies should treat this as a credential and lateral-movement risk vector.

Affected Systems

Windows, macOS, and Linux systems running Node.js/npm; developer workstations; CI/CD build environments; any AI coding agent or automated pipeline that installs npm dependencies without strict allowlisting

Indicators of Compromise

• Nearly 800 malicious npm package names (AI-generated/typo-squatted, specific list not disclosed in source)
• Cross-platform RAT payload (binary/hash not disclosed)
• Infostealer payload (binary/hash not disclosed)

Remediation Steps

1. 1

   Audit installed npm packages

   Review package.json and lockfiles across all projects for recently added or unfamiliar dependencies, especially those with names similar to popular packages.

2. 2

   Enable package allowlisting

   Restrict CI/CD and developer environments to use vetted internal package registries or allowlists rather than open npm resolution.

3. 3

   Scan for RAT/infostealer indicators

   Run endpoint detection tools to identify unauthorized remote access tools or credential-harvesting processes on affected hosts.

4. 4

   Rotate exposed credentials

   Immediately rotate API keys, cloud credentials, and secrets accessible from any system where suspicious packages were installed.

5. 5

   Restrict autonomous agent package installation

   For AI coding agents or automated pipelines that install npm dependencies, enforce human review or sandboxing before dependency resolution to prevent supply-chain compromise.

6. 6

   Monitor npm registry advisories

   Subscribe to npm security advisories and threat intelligence feeds like OpenSourceMalware to detect newly identified malicious packages promptly.