So I started with big expectation of what i can do with openclaw automation. However the product is not optimal for heavy automation.
Just keep that in mind. Personal assistant only.
So I started with big expectation of what i can do with openclaw automation. However the product is not optimal for heavy automation.
Just keep that in mind. Personal assistant only.
This is if u are stupid and don’t know how to code.
So i have been training this agent on openclaw and every time he makes mistakes I correct him and document the mistake. And he has been learning to build in openclaw.
The question is how to have his brain in the sub agents that he spawns with minimal token usage.
We can bootstrap all of this brain each time a sub spawn but this cost a lot. Specially with multiple agents spawning repetitively.
How would you solve for this?
Thanks in advance.
So what i built is a skill which is extremely useful and found great improvement with it.
The skill goes as follows:
When he makes a mistake I manually trigger a skill to make him capture it. The skill log the mistake in one of two documents:
1- universal: mistakes all agents learn from
2- local: mistakes only for him.
But spend so much time capturing the mistakes.
Do u have something similar?
Also if u figured how to automate this. Let me know.
Thanks
So, today we will say goodbye to 0auth ?
I’m new to this and have been extensively using my pro credit in claude for openclaw.
What are the agents that i should explore?
Budget 100$ usd per month
Usage: personal assistant