People on both sides are just getting started finding all the ways to abuse or protect you from security assumptions with these tools. RSS is the right tool for this problem and I would be surprised if their CMS doesn't produce a feed on its own.
I'm not totally naive. I had the VM fairly hardened originally, but it proved to be inconvenient. I relaxed it so that processes on the VM can see other devices on the network.
Now this tool spreads. You help everyone get it set up. Someone hacks the site, injects a prompt lying about some event, maybe Drag Queen Story Hour in a place with lots of people enraged about it. Now there's chaos and confusion. Corrections chase the spread of misinformation.
Giving plausible examples could further your case. But at some point you have got to realize that other people have actually thought about these things are are willing to do this.
Imagine going up to everyone riding a motorcycle and telling them about the inherent dangers of their activity and to stop. It is obvious that the OP understands risk, has taken several strong steps to harden their system and isn’t worried about the school calendar getting hacked making an event that they would get notified about and that destroying their community somehow. I don’t even understand openclaws place. The exact same events would unfold without the ai in there at all.
The trick is to give them a tree pruning chain saw, one intended for climbing tree loppers to use one handed - it's an ideal weight for nine years old to use two handed.
And to supervise.
As tested on my children and grand children.
Also, if you happen to have a furnace with a large pot of molten glass, five year olds are capable (given a stand) of making marbles from the furnance and will do that for hours if you can spare the time to let them.
Which is totally fine for the majority of tasks.
> Agents exfiltrate your data
They can only exfiltrate the data you give them. What's the worst that prompt injection attack will give them?