FakeGit Campaign Floods GitHub with 7,600 Fake Repositories to Distribute SmartLoader Malware
The FakeGit campaign has turned GitHub into a massive storefront of deceptive repositories aimed at distributing SmartLoader, a loader designed to install follow-on malware such as the StealC infostealer. The operation, active at least until July 2026, has created or repurposed nearly 7,600 repositories and employs a sophisticated mix of cloned legitimate projects, realistic documentation, and fake developer profiles to appear trustworthy.
Visitors are guided to download ZIP archives from GitHub Releases, a channel often perceived as more reliable. Approximately 200 of these repositories have recorded more than 14 million downloads, demonstrating the campaign’s significant reach. In roughly 800 cases the repositories masquerade as AI-related utilities, specifically presented as Skills or MCP (Model Context Protocol) servers.
This approach enables the AgentBaiting technique, in which attackers do not need to send direct links; instead, they rely on AI agents searching for free skills or MCP servers and subsequently following installation instructions that execute malicious code. Internal tests have shown that systems including Anthropic Claude Code, Google Gemini, and OpenAI ChatGPT can surface malicious repositories during discovery.
The infection chain begins with the downloaded ZIP, which triggers a staged payload executed through LuaJIT and culminating in an obfuscated Lua script that installs SmartLoader. The loader maintains persistence on Windows via scheduled tasks and resolves its command-and-control infrastructure dynamically using a blockchain-based dead drop referencing a smart contract on Polygon.
Exposure is further amplified by more than 600 listings appearing on public directories such as LobeHub, Glama, MCP.so, and MCP Market. Organizations are recommended to avoid executing unverified ZIP files or scripts, maintain internal catalogs of approved AI components, apply application control policies, and monitor for LuaJIT execution and anomalous network activity including direct IP connections or Polygon RPC endpoints.
Related articles
GitHub and PyPI Add Time-Based Defenses Against Supply Chain Attacks
GitHub and PyPI have introduced new time-based barriers to slow down supply chain attacks. Dependabot now waits a default of 72 hours before proposing version updates, while PyPI rejects new files added to releases older than 14 days. The changes target non-security version updates and attempts to poison older stable releases. Security updates remain immediate, and the cooldown can be adjusted via dependabot.yml. PyPI's restriction addresses risks from compromised tokens or CI/CD pipelines that allow malicious artifacts on past versions. The measures were implemented in July 2026 following incidents involving projects like LiteLLM and Telnyx.
GitHub and PyPI Introduce Time-Based Defenses Against Supply Chain Attacks
GitHub and PyPI have activated new time-based barriers to slow down supply chain attacks. Dependabot now waits a default of 72 hours before proposing version updates, while PyPI rejects new files added to releases older than 14 days. The changes target non-security version updates and attempts to poison older stable releases. Security updates remain immediate, and the cooldown can be adjusted via dependabot.yml. The PyPI restriction, effective since July 8 2026, addresses risks from compromised tokens or CI/CD pipelines. Both platforms aim to give the community time to detect malicious packages before widespread adoption.
Container Image Risks: Why Skipping Verification Today Breaks Your Service Tomorrow
Technical leader Nikita from Cloud.ru details four recurring container security failures observed across client environments. The article examines untracked vulnerabilities in high-privilege components such as the NVIDIA GPU operator, supply-chain compromises affecting even trusted tools like Trivy, persistent secret leakage patterns, and CI/CD pipeline exposure. Real incidents include a 2025 NVIDIA container toolkit exploit that enabled lateral movement to worker nodes and a 2026 Trivy GitHub compromise that poisoned over seventy image tags. The author stresses that pinning to image tags is insufficient and recommends full SHA256 hashes, private registries with caching, Cosign signatures validated by Kyverno or Connaisseur, and pre-commit secret scanning. Practical recommendations include automated CVE notifications and immediate patching of privileged components rather than waiting for sprint cycles.
Should Python Libraries Raise Minimum Dependency Versions to Block Vulnerable Releases?
Seth Larson of the Python Software Foundation argues that library maintainers should not automatically raise the minimum allowed version of a dependency after a vulnerability is disclosed. He states that dependency metadata exists to declare compatibility, while security decisions belong to the application owner who controls the final build. The recommendation has sparked debate in the Python community, with some maintainers supporting the separation of concerns and others viewing security constraints as part of library support obligations. CodeScoring’s analysis examines the practical impact on the ecosystem, noting that over 10,000 packages depend on urllib3 and tens of thousands more rely on numpy, requests, and pandas. The piece concludes that version bounds may incorporate security considerations only after evaluating real usage, affected versions, and downstream compatibility, but they cannot replace application-level vulnerability management with lockfiles and tools such as pip-audit.