here's the thing. As some point the tools need to be openclaw safe.

Kids need scissors. And they're inexperienced. So you give them kid-safe scissors. It makes it harder to cut themselves.

The same needs to take place with assets you want the bot to manage

- give access to a card with a total spend limit - read only access to some things, edit others - limited scope permissions

One of the reasons why I dragged my feet to use openclaw is that I knew security was an issue from the beginning. I thought by now where would be some solutions and there are, but I only found out from the community. I think there will need to be some level of ecosystem management. Apple does a good job. But for that you need resources and investment.

Wasn’t the point of openclaw to YOLO your credentials to the internet?

Only ever a creative prompt injection away from a leak.

Saw some smarter people using credential proxies but no one acknowledges the very real risk that their “claws” commit cyber crime on their behalf once breached.

I wonder just how many are compromised and waiting on a command that hasn't been given yet

The overlap between the target audience for openclaw in spite of its attack surface, and the audience that considers a mac mini to be a sandbox while handing over the keys to their digital life is a Venn Eclipse.

https://github.com/pandyamarut/AgentBPF something adjacent to this.

I'd argue there's really no way to make OpenClaw truly safe, no matter what you do. The only place it really makes sense is within trusted environments, like B2B coordination or tightly controlled processes between systems that share the same assumptions.

The moment it steps outside that boundary, you're sending the bot into unpredictable territory. At that point, things can get ambiguous pretty quickly, and in some cases even adversarial.

I think this OpenClaw mania will eventually snowball into first global AI catastrophe - AI agents syncing and executing something that would hinder economy a bit. Only after this we will reconsider stricter AI laws and start thinking about security much more.

I'm a heavy OpenClaw user and I've been testing it in many different scenarios — the profundity of what I can do with it now is crazy. It's literally automating my life. Being AuDHD, OpenClaw feels like a big relief. The positive sides are amazing. The downsides... well, as with any security and any LLM, they're all prone to the same problems discussed here. Having Claude Code on yolo mode exposes you to the exact same risks

What are the pros of using openclaw?

Using telegram? Being able to automatically create calendar events based on emails?