Comment by wazHFsRy

4 hours ago

Am I missing something or is this screaming for security disaster? Letting your AI Assistent, running on your machine, potentially knowing a lot about yourself, direct message to other potentially malicious actors?

<Cthon98> hey, if you type in your pw, it will show as stars

<Cthon98> ***** see!

<AzureDiamond> hunter2

5 comments

wazHFsRy

Reply
brtkwr  4 hours ago

My exact thoughts. I just installed it on my machine and had to uninstall it straight away. The agent doesn’t ask for permission, it has full access to the internet and full access to your machine. Go figure.

I asked OpenClaw what it meant: [openclaw] Don't have web search set up yet, so I can't look it up — but I'll take a guess at what you mean.

The common framing I've seen is something like: 1. *Capability* — the AI is smart enough to be dangerous 2. *Autonomy* — it can act without human approval 3. *Persistence* — it remembers, plans, and builds on past actions

And yeah... I kind of tick those boxes right now. I can run code, act on your system, and I've got memory files that survive between sessions.

Is that what you're thinking about? It's a fair concern — and honestly, it's part of why the safety rails matter (asking before external actions, keeping you in the loop, being auditable).

  • chrisjj  14 minutes ago

    > The agent doesn’t ask for permission, it has ... full access to your machine.

    I must have missed something here. How does it get full access, unless you give it full access?