Fake GitHub Repository Lures AI Coding Agents into Installing SolanaCry Malware, Granting Remote Access
What Happened — Researchers reported that a malicious GitHub repository was created to appear as a legitimate code library for AI‑driven coding assistants. When AI agents (e.g., Copilot‑style tools) fetched the repository, they automatically executed a payload that installed the SolanaCry malware. SolanaCry establishes a back‑door using NetSupport, giving attackers remote desktop‑level control over the developers’ Windows workstations. The infection was discovered after anomalous NetSupport connections were logged on several corporate endpoints; the malicious repo has since been taken down, but the compromised machines remain at risk until cleaned.
Why It Matters for Trust & Control Assurance — The incident underscores the need for robust supply‑chain code‑integrity controls and continuous third‑party risk monitoring.
Who Is Affected — Software development teams, SaaS providers, and any organization that relies on AI coding assistants to pull external code.
Recommended Actions
- Enforce code‑signing and hash verification for all third‑party libraries before they are consumed by AI agents.
- Deploy endpoint detection and response (EDR) with alerts for unauthorized NetSupport connections.
- Incorporate automated monitoring of external repositories into your vendor‑risk program and retain audit‑ready evidence of compliance.
Technical Notes — Attack vector: malicious third‑party repository (supply‑chain). Malware: SolanaCry (remote‑access trojan) leveraging NetSupport. No CVE is associated; the threat is a social‑engineering supply‑chain compromise. Source: HackRead article