The point: The FakeGit campaign exploits counterfeit GitHub repositories with copied projects and deceptively authentic developer profiles to distribute SmartLoader malware through fake AI tools and MCP servers.
Security researchers have identified nearly 7,600 malicious GitHub repositories, with over 800 of them impersonating AI capabilities or Model-Context-Protocol servers to distribute the SmartLoader malware family.
Security researchers have uncovered a large-scale campaign called FakeGit that encompasses nearly 7,600 malicious GitHub repositories. More than 800 of these repositories claim to provide AI capabilities or Model-Context-Protocol servers. In reality, these repositories serve to distribute the SmartLoader malware family.
The campaign employs multiple obfuscation techniques: copied or forged open-source projects, deceptively authentic developer profiles, professionally designed README files, and manipulated ZIP archives. This combination is designed to deceive developers searching for legitimate AI tools or MCP implementations, particularly those intended for agentic AI applications or Claude integration.
For CISOs, this campaign is relevant because it opens a new attack surface against development teams and their supply chain. By targeting popular development platforms and masking itself as trusted AI components, FakeGit evades both technical control mechanisms and developer awareness. The discovery of nearly 7,600 repositories indicates a systematic and well-organized operation.
It is recommended to review dependency management, train developers on AI tool verification, and critically validate repository downloads and external dependencies. A focus should be placed on verifying MCP server sources, especially when deployed in automated or agentic AI environments.
Source: thehackernews.com · Published 20 July 2026
Lumi AI News — AI-assisted curation pursuant to Article 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.7.3.