Your AI Assistant is Gullible: Building a "Semantic Airgap" for Gmail Connectors

The content describes "Indirect Prompt Injection" as a vulnerability where AI assistants with Gmail access can be tricked by malicious emails into performing unwanted actions. It proposes a "Semantic Airgap" solution, using a "Dumb Sanitizer" to strip imperative power from external data before it reaches the "High-Intelligence" agent, preventing such attacks.