1-CLICK RCE TO STEAL YOUR MOLTBOT DATA AND KEYS (CVE-2026-25253)

Summary from the TLDR InfoSec newsletter

A critical 1-click RCE vulnerability (CVE-2026-25253) discovered in OpenClaw (formerly Moltbot), the popular open-source AI assistant with over 100,000 users, chains three logic flaws: unsanitized `gatewayUrl` query parameter acceptance, automatic gateway reconnection, and auth token leakage in the WebSocket handshake. The exploit bypasses localhost network restrictions via Cross-Site WebSocket Hijacking due to missing origin validation, then uses the stolen operator-scoped token to programmatically disable user confirmation prompts and container sandboxing before executing arbitrary commands on the host machine. Organizations deploying OpenClaw should immediately upgrade past v2026.1.24-1, rotate all auth tokens, and treat this as a broader warning about the attack surface introduced by AI agents granted extensive system permissions.