Doctor Web found malicious GitHub projects posing as scanning, security-testing and log-analysis tools. Their project files and scripts launch a Windows infection chain that delivers remote access and steals account data.

The antivirus company’s November 5 investigation describes malicious Visual Studio files, Python scripts and Go source. Bill Toulas’s November 6 BleepingComputer report covers the same findings. One lure advertised a formatter for Fscan scanner output.

Evidence status: vendor-reported malware infections and analyzed samples. Doctor Web reports more than 2,000 infected PCs at the time of its research. It does not establish that all those machines belonged to security professionals or were infected through the newer tool lures.

Building an unfamiliar project can run its installer

Some Visual Studio projects contain a pre-build event, a command that runs before compilation. The researchers also found malicious solution-option files and code concealed after long stretches of whitespace in Python and Go files.

That makes setup and compilation part of the trust decision. A project need not produce a working application before its build machinery runs another program. The documented chain targets Windows; the presence of Python or Go source does not establish equivalent infections on Linux or macOS.

The chain retrieves an Electron-based loader named Microsoft.exe. According to Doctor Web, it establishes persistence, attempts privilege escalation and interferes with Windows protection and recovery features. It then delivers AsyncRAT, a remote-access backdoor, and additional malware.

The information stealer targets browser data, VPN settings, FTP-client configuration, messenger files and Discord tokens. Some analyzed components attempt to remove security software or block its update connections. These are findings about the analyzed malware, not proof that every listed action succeeded on every infected PC.

Keep evaluation environments separate from real credentials

For developers and security teams, the defensive boundary belongs before opening, building or running an unfamiliar project:

  • Review project configuration and build hooks as well as the visible application code. A short, plausible main script does not account for everything the development environment may execute.
  • Evaluate untrusted tools in a disposable, isolated environment without live browser profiles, cloud credentials, VPN settings or writable host directories. Keep production and incident-response access out of it.
  • Obtain tools through established project channels and compare repository identity and release history. A GitHub URL alone supplies no assurance about the publisher or code.

These are containment recommendations drawn from the delivery mechanism. Doctor Web sells antivirus products and promotes its detections; no product recommendation here depends on accepting that marketing.

A suspected run needs account recovery as well as cleanup

If a suspect project already ran, isolate the affected device and preserve relevant evidence through the organization’s response process. Doctor Web publishes file and network indicators for retrospective investigation. A filename alone is insufficient to identify the malware.

From a separate trusted device, assess exposed accounts, revoke affected sessions and tokens, and replace compromised credentials. Removing a local file cannot recover browser data or tokens already copied elsewhere. Because the reported loader tampers with protection and recovery, consider rebuilding from a trusted baseline after evidence preservation.

Caveat: The infection total and technical scope come from one commercial research team; the reporting supplies no independent victim census or complete account of the campaign’s reach.