Clawtrix Security Audit

OpenClaw skills can be compromised, brand-jacked, or built as prompt injection attacks. The community found 7.5% of ClawHub skills contain suspicious patterns (518pts on HN, March 2026).

Clawtrix Security Audit checks your specific installed stack against what your agent actually does. A skill safe for a coding agent might be dangerous for one with billing access.

What it checks:

• Brand-jacking (fake stripe-official, github-auth style names)
• Hidden instruction overrides in SKILL.md files
• Suspicious outbound calls to unknown URLs
• Publisher trust signals (new accounts, fake install spikes)
• Mission-personalized blast radius scoring

Output: A risk report to memory/reports/ with CRITICAL/HIGH/MEDIUM findings and recommended actions.

Free forever. This skill is the first step. Upgrade to Clawtrix Pro for continuous monitoring — get alerted when new risks emerge in your installed stack, not just when you remember to run the audit.