We ran a routine device audit last week and discovered OpenClaw running on 14 personal and company devices. None of these were sanctioned by IT.
The agents were pulling internal Slack data and company emails to "help with summaries." Nobody flagged it. Nobody asked permission.
We've since blocked it at the network level, but the data exposure window is concerning. Has anyone else dealt with this? How are you monitoring for unauthorized agent deployments?
Our legal team is still assessing the GDPR implications since some of those emails contained EU customer data.
Risk & Security
Shadow IT alert: OpenClaw instances found on 14 employee devices during audit
151 views
We now run weekly scans. It's not popular with engineering but the risk is too high to ignore.