Fast-Growing Open-Source AI Assistant Is Testing the Limits of Automation—and Safety
Clawdbot, an open-source AI assistant, has rapidly gained popularity among developers, accumulating over 10,200 GitHub stars and 8,900 Discord members since January. Unlike mainstream assistants, Clawdbot executes real-world tasks autonomously—such as making restaurant reservations by phone when online platforms fail—and works across messaging apps like WhatsApp, Telegram, Discord, Slack, Signal, and iMessage. It operates locally, supports persistent memory, can execute terminal commands, control browsers, manage files, and connect to AI models (e.g., Claude, GPT) via the Model Context Protocol. The platform’s user-friendly installation and fast development have fueled its growth, but serious security concerns have emerged: by default, many installations expose gateways without authentication, leaving devices vulnerable to attacks. The suggested fix involves restricting network access and securing configurations, though adoption of safeguards lags. Users are also encountering unexpectedly high API token fees, with some spending hundreds of dollars in days. Created by Peter Steinberger, Clawdbot presents impressive capabilities, but users are advised to secure installations and monitor costs closely.
