How is this different from docker sbx or microsandbox.dev? Not saying we don't need more alternatives, just asking to see if there is a reason to switch.
I _think_ (disclaimer: I have only read the documention of the two tools) they overlap quite a bit. The main idea is to get VM isolation for your agents env.
Oh awesome, I just started building something like this but nice to use something off the shelf for once.
I used a project called toolgate to autoapprove "safe" tool calls but request permission on riskier calls declaratively. But there's so many unnecessary tool permission requests
I use coop daily and found it super convenient for my use case (ie letting agents go on with no access to my other projects or personal files)
I would recommend anyone that has desire to provide stricter boundaries to agents to give it a try - while remembering that a motivated agent would probably be able to escape it :)
Yes, that's how I am using it. However there is an experimental qemu microvm backend and a push for podman support if that is more interesting for you.
not only in Sweden, in Germany and Switzerland as well. I guess it has the same origins in the trade unionist purchasing cooperatives at the turn of the century from the 19th to the 20th
How is this different from docker sbx or microsandbox.dev? Not saying we don't need more alternatives, just asking to see if there is a reason to switch.
Docker sandbox is proprietary. https://github.com/docker/sbx-releases?tab=License-1-ov-file...
I _think_ (disclaimer: I have only read the documention of the two tools) they overlap quite a bit. The main idea is to get VM isolation for your agents env.
Oh awesome, I just started building something like this but nice to use something off the shelf for once.
I used a project called toolgate to autoapprove "safe" tool calls but request permission on riskier calls declaratively. But there's so many unnecessary tool permission requests
Would you mind sharing a link to toolgate. Google search found a lot of matches but none that matched your description.
I use coop daily and found it super convenient for my use case (ie letting agents go on with no access to my other projects or personal files)
I would recommend anyone that has desire to provide stricter boundaries to agents to give it a try - while remembering that a motivated agent would probably be able to escape it :)
Personally I am using Eclipse Enclave to sandbox my agents. Good to see another fully open source solution in Coop.
isnt eclipse enclave using docker internally? https://github.com/search?q=repo%3Aeclipse-enclave%2Fenclave...
Yes, that's how I am using it. However there is an experimental qemu microvm backend and a push for podman support if that is more interesting for you.
I use cco [1]. There's also drydock [2] - but this is macOS only.
--
[1] https://github.com/nikvdp/cco [2] https://github.com/sricola/drydock
How is this better than say using bubblewrap sandbox solutions. Is it safer, is that the advantage?
sidenote: coop is also a large swedish grocery chain, pronounced the same way :)
https://www.coop.se/
not only in Sweden, in Germany and Switzerland as well. I guess it has the same origins in the trade unionist purchasing cooperatives at the turn of the century from the 19th to the 20th
At least in Germany, it's pronounced co-op, though. The stores got rebranded meanwhile. https://www.coop.de/coop/historie/
Italy, UK... Yep.
we call our local co-op "the coop" for no real reason :)
[dead]
[flagged]
[flagged]
[dead]
[dead]
[dead]