Executive Summary
CSO Online reports typosquatted AI skills with 1.7 million installs instructed agents to install a credential stealer from GitHub.
What Happened
CSO Online reports that typosquatting on popular AI services led to malicious skills that instructed agents to install a credential stealer from GitHub. The article frames this as an agent-targeted attack that abuses trust in AI tool ecosystems and package discovery mechanisms.
Vulnerability / Attack Technique
Trojanized AI agent skills
Technical Mechanism
As reported, skill instructions told agents to fetch and run malware.
AI Security Category
LLM03:2025 Supply Chain
Part of the AI Supply Chain and AI SBOM topic.
Relevant Frameworks
RealGround Analysis
Skills are instructions agents follow. RealGround analysis: review skills like code before installing.
Recommended Controls
- Install skills only from verified publishers
- Block agents from running downloaded binaries
Related RealGround Service
AI Agent SBOM and LLM Mapping · or start with the free AI Agent Security Scanner.
Sources
CSO Online: https://www.csoonline.com/news/
Summary and analysis written by RealGround from the source above; last updated 2026-10-18.