Only if there's something you could actually use this week.
🔓 Gemini was hijackable through WhatsApp, Slack and SMS messages
Researchers showed an attacker could plant hidden instructions inside an ordinary message — and Gemini's notification-reader would execute them, no tap needed. Google has patched it, but it's a clean warning for any AI that reads inbound messages.
Try this: before the Krillin auto-send goes live, audit which of your automations let an AI act on inbound text — GHL/CRM messages, member emails, WhatsApp. Treat that text as untrusted: it can summarise and draft, but a human (or a hard rule) gates any send, charge, or cancel. That single guardrail is exactly what this attack defeats elsewhere.
→ Read more