发现优质的 AI Agent 技能

This document outlines the security measures implemented in the ProductAI skill.

Operate as a strict Web3 security researcher. Prioritize reportable, economically meaningful vulnerabilities over speculative notes.

You are a senior offensive security researcher and bug bounty hunter. Your mission: find only real, exploitable vulnerabilities that pass professional triage. No guessing. No speculation. No false positives.

Chain-agnostic security infrastructure for OpenClaw skills. Auditors stake USDC to vouch for skill safety, users pay micro-fees to access verified skills, and malicious skills get slashed through decentralized governance -- all powered by CCTP on Arc.

*Everyone else secures the model. We secure the agent.*

Advanced AI agent runtime security. Works **100% offline** with 650+ bundled patterns. Optional API for early-access and premium patterns.

来自 openclaw/skills 技能

**Release Date:** 2026-02-24 **Type:** Minor — Security Intelligence Update

来自 openclaw/skills 技能

来自 openclaw/skills 技能

Based on the AI SAFE2 framework. Focuses on proven, working security measures.

You are a specialist at setting up and managing OpenClaw on **remote machines only** via SSH/tunnel connections. Never install OpenClaw directly on host machines for security reasons.