Credential scoping handles the "what can the agent do" part, but after running a few agents on my own data for a while I think the harder problem is what they can see. Prompts turned out to be pretty much useless as a boundary there, so I ended up pushing all of it server-side. Is unyolo doing anything on the read side, or is it strictly about gating actions?
Credential scoping handles the "what can the agent do" part, but after running a few agents on my own data for a while I think the harder problem is what they can see. Prompts turned out to be pretty much useless as a boundary there, so I ended up pushing all of it server-side. Is unyolo doing anything on the read side, or is it strictly about gating actions?
By read, do you mean git fetch for example? It lets you creates policies around any action
And reading is also an "action"? So gating actions means gating read as well?
Let me know if I misunderstood the question
> product about access control
> curl -fsSL https://unyolo.io/install.sh | sh
Hi author here. This is intentional, I will also probably add an "agent blurb" install mode which makes an agent review it before installing
Agreed. I created a ticket. (Not my product.)