AI Infra Guard

AI Infra Guard is a self-hosted AI security testing platform for developers and security teams checking agent workflows, model responses and AI infrastructure. It combines vulnerability scans and adversarial prompt tests in a web interface.

• Agent and OpenClaw checks: Test Dify and Coze workflows for OWASP Top 10 for Agentic Apps risks, or request OpenClaw assessments through chat for insecure settings, risky skills, known vulnerabilities and privacy leakage.
• MCP and skill audits: Scan source code or remote URLs for tool poisoning and code vulnerabilities; skill checks also cover memory poisoning, unauthorized access and insecure dependencies, with CI/CD integration.
• Infrastructure scans: Identify AI components and check their web services for known CVEs, including Ollama, ComfyUI, vLLM, n8n and Triton Inference Server.
• LLM jailbreak evaluation: Compare models' responses to malicious, adversarial and deceptive prompts, then export failed cases for guardrail work.

The software is free and open source under Apache 2.0. It runs through Docker on Linux, macOS and Windows, with at least 4 GB of RAM and 10 GB of disk space. A separate online offering also exists.

The interface supports English and Chinese, and an API lets other applications request scans. Plugins extend component detection, vulnerability checks and evaluation datasets. A model and API relay checker supports model fingerprinting and relay auditing; its frontend runs separately.