This is a security scanner designed to analyze AI agent skills (OpenAI Codex Skills, Cursor Agent Skills) for threats including prompt injection, data exfiltration, and malicious code. It combines pattern-based detection using YAML and YARA rules with LLM-as-a-judge evaluation and behavioral dataflow analysis to identify potential security issues. The tool scans agent skill definitions and associated code to flag suspicious patterns and help developers harden their AI agent implementations.