← Back to talks

Talk · AWS User Group Security Ecuador

What permissions do you give an AI agent on EKS and AWS?

Three questions, with live demos on a shop running on EKS: how to keep an agent from reading sensitive data or changing something nobody asked for, what stops it when someone gives it orders through a log line, and how to give it Bedrock models without anything leaving AWS.

Online Given in Spanish, slides translated 31 slides

Click the slides and use the arrow keys, or tap either side of the slide. Open in a new tab Watch the recording (in Spanish)

Takeaways

If you're going to give an agent access

  1. Put the limits where it executes In the gateway, the session and the role. The prompt helps, but it doesn't protect.
  2. Redact before it reads What the model never sees, it can't leak, not even when tricked.
  3. Let it propose, and a person execute With their role, and with the server checking again.
  4. Find the outbound channel If there's no way to get something out, an injection wins nothing. In our case, it showed up in an image.

← Back to talks