🔒 Security

Agent Security Hardener

Runtime security monitoring and threat detection for OpenClaw agents. Continuous protection against prompt injection and data exfiltration.

What it does

Agent Security Hardener monitors every incoming message and outgoing action in real time, scanning for prompt injection, tokenade attacks, siege patterns, wallet drain attempts, and data exfiltration. When a threat is detected, it auto-quarantines the suspicious skill or session and alerts you immediately.

Unlike security-hardening-wizard (a one-time audit), this skill provides continuous runtime protection — a security camera that never sleeps. It tracks plugin provenance, monitors rate anomalies, and generates real-time security alerts with severity scores.

Use cases

• Production agent deployments exposed to untrusted users or external inputs
• Teams that install third-party skills and want supply chain protection
• Financial agents handling real transactions that need exfiltration detection
• Enterprise deployments requiring continuous security audit logging

Requirements

• OpenClaw 3.20+
• Python 3.10+

How to install

After purchase you’ll receive a download link by email. Extract and copy the agent-security-hardener/ folder into your OpenClaw skills/ directory, then restart OpenClaw.