Skill Injection — attacks against LLM agent skill ecosystems
Skill Injection is a security-focused resource that documents how malicious agent skills can be inserted into LLM agent ecosystems, highlighting the mechanisms, risks, and defensive considerations. The site targets researchers, developers, and security professionals seeking awareness of this evolving attack class and its implications for trusted capabilities and supply-chain trust in AI-powered agents.
We analyze text and images across millions of websites using similarity detection to identify large-scale content duplication and reuse.
This signal indicates whether widespread duplication patterns were detected at the time of analysis.
Signals update over time.
Discussion
Log in to comment or vote.