Description
An agent skill installed from the web may hide malicious instructions or dangerous scripts you cannot spot beforehand. SkillWard is a security scanner for AI agent skills that uncovers hidden threats before deployment.
It combines static analysis, LLM evaluation and sandbox verification to find risks.
Static analysis:Suspicious code.
LLM review:Malicious instructions.
Sandbox:Isolated verification.
It combines static analysis, LLM evaluation and sandbox verification to find risks.
Features
Static analysis:Suspicious code.
LLM review:Malicious instructions.
Sandbox:Isolated verification.
