Fake Web3 Accelerator Distributes Malware Through npm Packages Promising Free AI Credits

A counterfeit Web3 accelerator has been identified as the source of malicious npm packages that install information-stealing software and backdoors on Windows and macOS computers. The scheme lures developers by advertising complimentary credits for popular artificial‑intelligence services such as Anthropic Claude and Google Gemini, as well as access to coding tools.

Security researchers discovered that the bogus accelerator publishes a series of npm modules that appear legitimate at first glance. When developers add these packages to their projects, the code silently drops payloads capable of exfiltrating credentials, browsing histories, and other sensitive data. In addition, the malware creates a foothold that can be used to commandeer traffic generated by AI agents, potentially redirecting queries or injecting malicious responses.

The operation exploits the trust inherent in the npm ecosystem, where open‑source libraries are frequently reused across countless applications. By bundling the malicious code within packages that claim to simplify integration with AI APIs, the attackers tap into the current surge of interest in generative AI. The promise of “free AI credits” is a particularly effective bait for independent developers and small startups eager to experiment without incurring cloud‑service costs.

While the exact origins of the fake accelerator remain unclear, the tactics mirror previous supply‑chain attacks that have targeted npm, such as the infamous event‑stream incident. Researchers note that the malicious packages are signed with valid npm accounts, making them harder to flag automatically. Users who install the modules without scrutinizing the source or reviewing the code are at risk of having their systems compromised.

Cyber‑security experts advise developers to verify the authenticity of any third‑party libraries, especially those that claim to provide free credits or exclusive tooling. Checking package provenance, reading community reviews, and employing automated scanning tools can help detect hidden payloads before they are executed. Organizations should also enforce policies that limit the use of unvetted npm packages in production environments.

The discovery underscores the broader challenge of securing the software supply chain as AI services become more pervasive. As the demand for rapid AI integration grows, threat actors are likely to continue masquerading as helpful platforms. Ongoing monitoring of npm repositories and swift coordination between security researchers, platform providers, and the developer community will be essential to mitigate future attacks.