3 ms·
An agent should never have access to files or data outside of the project (emails, passwords, maybe crypto, photographs) - it makes no sense to allow that. I'v
by jeswin 1y ago
An agent should never have access to files or data outside of the project (emails, passwords, maybe crypto, photographs) - it makes no sense to allow that.
I've always run agents inside a docker sandbox. Made a tool for this called codebox [1]. You can create a docker container which has the tools that the agent needs (compilers, test suites etc), and expose just your project directory to the agent. It can also bind to an existing container/docker-compose if you have a more complex dev environment that is started externally.
[1]: https://github.com/codespin-ai/codebox/ https://github.com/codespin-ai/codebox/
- hoherd 1y agoThere is also https://containers.dev https://containers.dev which is the first thing I configured when my company gave us Cursor. All agentic LLMs to date need to be on a very short leash. Pro tip: you can run `docker diff <container-id>` to see what files have changed in the container since it was created, which can help diagnose unexpected state created by the LLM or anything else.