Prompt Injection Defense

Use when an AI agent or LLM application consumes untrusted user/retrieved content, uses tools or connected data, and needs defenses against direct/indirect prompt injection, data exfiltration, instruction hijacking or unsafe tool execution.