I begin with saying sorry if you idea is a bit wild but I come for a different "sandbox" than development (old school on-prem+cloud+infrastructure).

What about having a Linux container that had strands-agents SDK with some tools like "Utilities", "Web & Network", "Code Interpretation", "Shell & System" and "File Operations" enabled by default. Add some standard model providers and option to change between bedrock or local storage, wire all the configs into environmental variables to customize at build.

My vision is to have an isolated Linux container (wsl) with a strands-agent + tools enabled and integrated to the OS with essentially full control. How well would it work? How much could the agent "grow" with use of memory etc. could it become increasingly better? how complex tasks could you get it to do?

What do you think, am I crazy or could this work?