Introducing SandyClaw - The First Dynamic Sandbox for AI Agent Skills and Prompts

### Executive summary This report introduces SandyClaw, a Permiso Security platform that applies dynamic sandbox detonation to AI agent skills by executing skills in an instrumented environment to capture full LLM and OS-level behavior, enabling detection of malicious skills (credential stealers, reverse shells, data exfiltration) that static or LLM-based scanners miss; it emphasizes multi-engine analysis, SSL interception, and a searchable intelligence repository to help security teams validate behavior-based verdicts.