XClaw - Security-hardened self-hosted AI agent with one-line install

XClaw is a security-hardened fork of OpenClaw — the open-source self-hosted AI agent framework. What makes XClaw different: • 7 new security scanning rules (prompt injection, reverse shell blocking, token theft detection, data exfiltration protection) • Plugin filesystem permission guard • Hardened systemd service MIT licensed. Your server, your data, your control.

Hey Product Hunt! 👋 I built XClaw after running OpenClaw for a few months and noticing some serious security gaps in the skill execution pipeline. XClaw adds military-grade hardening on top: - Blocks prompt injection attempts from malicious skills - Detects reverse shell patterns (bash -i, nc -e, etc.) - Prevents API key and gateway token leaks - Stops data exfiltration via curl The one-liner installs everything — Node.js, nginx, SSL, systemd service — on any Debian/Ubuntu server. Then pair it with Telegram and you have your own AI agent running in minutes. Happy to answer any questions! 🦞

XClaw - Security-hardened self-hosted AI agent with one-line install

Replies