agentguard
Real-time security scanning and runtime protection for coding agents
All Skills in This Repository (1)
Pure Emerald Level IndicatorsFrequently Asked Questions
FAQPage SchemaHow to install AgentGuard?▼
Run `npx skills add GoPlusSecurity/agentguard --all -g -y` in your terminal to install the skill globally, then run `agentguard init` to configure hooks for your agent.
How to protect my AI agent from malicious skills?▼
AgentGuard auto-scans every new skill on session start with 24 detection rules and blocks dangerous code before it runs. It also tracks which skill initiated each action so malicious skills are held accountable.
Can AgentGuard block dangerous terminal commands?▼
Yes. Its runtime guard blocks destructive commands like rm -rf /, fork bombs, and curl-pipe-bash, and prevents writes to credential files like .env and .ssh/.
Does AgentGuard work with Claude Code and OpenClaw?▼
Yes. It supports Claude Code, Codex, OpenClaw, Hermes, QClaw, and DSH through native hooks, plugins, and the universal SKILL.md standard.
Does AgentGuard detect Web3 and crypto threats?▼
Yes. It detects wallet draining, unlimited token approvals, reentrancy, proxy exploits, and private key leaks, with optional GoPlus API credentials for transaction simulation.
Related Repositories in Legal & Compliance
View All in Legal & Compliance→claude-for-legal
AI legal workspace for contracts, privacy, IP, litigation, and compliance
patent-disclosure-skill
Draft Chinese patent disclosures and read patents in plain language
app-privacy-policy-generator
Generate privacy policies and terms for mobile and web apps