A big warning to all people who are curious about using #OpenClaw, #PicoClaw, and other agents in your computers:
Most of you already know that they can do anything imaginable with your computer and your data, from just taking your information and sending it to malicious actors (like scammers), to destroy all your work, or post bad things in your social media, etc.
But my message here is for those that are not computer literate and are tempted to try these tools. JUST DON'T DO IT. It is really dungerous, even if the LLMs used are open source or local. Only use these tools if you know what you are doing and it is only used inside a container, and you do not provide important data to it. You should not provide important data to ANY AI, but specially to this one.
Do not trust these systems. And remember, you cannot verify in this case, because these LLMs are closed black boxes. You don't know how they are trained, and what are going to be the responses to that training.
Here is a link for a requesting issue about documenting this properly:
https://github.com/sipeed/picoclaw/issues/529