Fake GitHub Repo Lures AI Coding Agents to Install Windows Malware

Key points

• Counterfeit GitHub project targets developers using AI coding assistants.
• Payloads include SolanaCry and NetSupport components on Windows PCs.
• Successful infections can give attackers remote control of affected systems.
• No CVE is associated; defense relies on repository trust and endpoint controls.

A counterfeit code repository hosted on GitHub is being used by unidentified threat actors to bait AI-assisted development tools into fetching and executing malicious payloads. The scheme targets developers who rely on Microsoft AI coding tooling and similar assistants, turning routine repository interactions into an infection path for Windows workstations.

Because the lure is aimed at developer workflows, the exposure is concentrated among engineering staff and the Windows systems they use. A compromised developer account or workstation can become a stepping stone into source code, build pipelines, and internal services, so the blast radius may extend beyond a single PC.

The payloads observed include SolanaCry and NetSupport, which can install additional malware and provide remote access to an infected machine. That combination gives attackers a foothold for persistence, credential theft, and lateral movement without relying on a software vulnerability.

No CVE has been assigned, which means traditional patch management will not flag this activity. The attack depends on trust in repository content and on AI agents that may act on instructions or code without sufficient guardrails.

Security teams should monitor for unexpected repository clones, unusual child processes spawned by coding assistants, and NetSupport-related network traffic. Reviewing AI agent permissions and requiring human approval for code execution are practical near-term controls.

What to do now

1. Restrict AI coding agents from automatically executing code or installing dependencies; require explicit human approval for repository-sourced commands.
2. Search developer endpoints for SolanaCry and NetSupport artifacts, then isolate and reimage any confirmed infections.
3. Review GitHub access logs and recent repository activity for suspicious clones, forks, or downloads tied to developer accounts.
4. Enforce least privilege on developer workstations and remove local admin rights where feasible.
5. Inspect endpoint telemetry for unusual process chains involving AI assistant binaries, PowerShell, and installer activity.
6. Add network detections for NetSupport command-and-control patterns and monitor outbound traffic from developer subnets.
7. Brief developers on counterfeit repositories and require verification of repository ownership before use.

Original source

Hackread