An open-source AI agent called OpenClaw (formerly known as both Clawdbot and Moltbot) that runs on your own computer and “actually does things” is taking off inside tech circles. Users interact with OpenClaw via messaging apps like WhatsApp, Telegram, Signal, Discord, and iMessage, giving it the keys to operate independently, managing reminders, writing emails, or buying tickets.  But once users give it access to their entire computer and accounts, a configuration error or security flaw could be catastrophic. A cybersecurity researcher also found that some configurations left private messages, account credentials, and API keys linked to OpenClaw exposed on the web. Despite the potential risk...