Awesome AI Agents Security
A curated list of open-source tools and frameworks for securing autonomous AI agents, organized by security lifecycle stage from red teaming through runtime protection to governance.
Awesome AI Agents Security is a curated list of open-source tools, frameworks and resources for securing autonomous AI agents, organized by the security lifecycle of an agent rather than alphabetically: agent firewalls and gateways for runtime protection, red-teaming and vulnerability scanners, static analysis and linters, sandboxing and isolation environments, guardrails and compliance, benchmarks and datasets, and identity and authentication.
Listed runtime-protection tools include an agent-native proxy adding role-based access control and policy enforcement to agent-tool interactions, and a security-focused agent runtime that scans for prompt injection, MCP risks and unsafe package installs before execution. Listed red-teaming and vulnerability-scanning tools include an autonomous penetration-testing agent run inside a Docker sandbox, a multi-turn adversarial attack framework, a fuzzing scanner for agent workflows, a vulnerability scanner for model hallucination and prompt-injection susceptibility, and a scanner for agent-to-agent protocol communication.