Hello, My name is Jad and I am building an openclaw agentic framework to automate what I can from my e-commerce business. Looking forward to meet all of you
@Raul Flores yes the memory is a governed shared memory layer that i created, memory doesn’t get created randomly, it needs to be reviewed by the memory agent that forward it to me to be patched with a script that i run. This way memory never has conflicting info or bloating for no reason. Memory also contain some behaviorof agents that i can broadcast with my script and everyone inherits.
@Raul Flores The weak point is mostly in openclaw and what it allows you when you want to call sub-agents and when you want agents to call other agents. All of this is especially complicated with sandboxing, my current understanding is that my architecture can never be done with a sandbox. openclaw is too technical and you really need to understand every lever, otherwise you will have silent failures.