Local AI Agent Runtime Protection

Inspects the local agent loop on endpoints — user prompts, tool requests before execution, and tool responses — to detect and block prompt injection and high-risk actions by CLI tools, coding assistants, and desktop AI apps.