Security Skills for AI Agents

Security skills help agents check what they are about to touch: verifying URLs before a browser visit, auditing a gateway host, and managing credentials without leaking them into model context. This matters more as agents act autonomously. OpenClaw skills distributed through ClawHub are scanned against VirusTotal, which lowers supply-chain risk but is not a guarantee, so read the SKILL.md of anything that wants shell access or API keys.

Other skill categories