Open Source AI Agent Security Projects
Tools that protect AI agents, their tools, MCP servers, plugins, and package installs from prompt injection, secret exposure, unsafe commands, and supply-chain attacks at runtime.

HOL Guard
Local-first antivirus for AI agents that blocks risky commands, secret access, prompt injection, malicious packages, and MCP threats at runtime.
Alternative to:
HiddenLayer
HiddenLayerAI-assisted

OpenShell
NVIDIA's open-source runtime that runs AI agents in kernel-isolated sandboxes with policy-enforced file, network, and credential access.
Alternative to:
E2B
E2B