cat /playbooks/agent-security --newest-first
Securing agents and AI tools
Sandboxing, scoped credentials, prompt-injection defence, permission and approval gates, egress control, supply chain of models and tools.
recommendations 0
No recommendations collected for this problem yet.