infosec

"Instead of creating ‘AI agent’ permissions, fix your existing ones … You need separate read and write permissions for each action, and you need to clearly document what each permission allows."

"Instead of creating ‘AI agent’ permissions, fix your existing ones … You need separate read and write permissions for each action, and you need to clearly document what each permission allows."
https://mayakaczorowski.com/blogs/ai-agent-authentication

security engineering and applied cryptography mantej.blog

This is literally it. In most cases, the AI agent is a proxy for a human user.

So just use your existing permissions engine