Security flaws in agent infrastructure from Amazon Web Services AWS, Google, and Vercel let untrusted or forged instructions reach an agent's tools with no check that a model turn had authorized them. In several of the attack paths, the model never ran at all, so system prompts, content filters, and modellevel guardrails never got a chance to intervene. The affected products include Amazon.📖 Read more.
🔗 Via "The Hacker News"
----------
👁️ Seen on @cibsecurity