Deploy Now

SaaS Alternative

HiddenLayer Mindgard Lakera Red SPLX

Stars

7,587

Forks

780

Watchers

50

Developer links

AI-Infra-Guard

From MCP servers and agent skills to Ollama endpoints and LLM guardrails, AI-Infra-Guard probes the whole AI stack for security holes from one chat-style web console. Infra Scan fingerprints 100+ AI frameworks such as Ollama, ComfyUI, vLLM, Gradio, and Triton, then matches versions against 2,000+ known CVEs for single targets or uploaded batch lists, with port scanning for bare IPs. MCP Scan audits a server from a zip file, a GitHub repository, or a live SSE or streamable HTTP address, hunting for tool poisoning, rug pulls, data theft, indirect prompt injection, and plaintext keys. Skill Scan puts agent skills through the same three stages of information collection, code audit, and vulnerability review. Agent Scan tests Dify, Coze, and HTTP chat agents against the OWASP Top 10 for agentic apps, and Jailbreak Evaluation fires curated datasets and attack methods at any OpenAI-compatible model, then exports the failed cases. Reports give a score, a security level, evidence, an attack path, and fixes, while progress streams live as the scanning agents work. Plugin management edits fingerprint rules, vulnerability entries, evaluation datasets, and MCP plugins right in the browser. You bring your own LLM key for the scanning agents. There is no built-in login, so keep it on a private network or behind a reverse proxy. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

AI-Infra-Guard
AI-Infra-Guard
AI-Infra-Guard
AI-Infra-Guard
AI-Infra-Guard

Benefits

  • Audit The Whole AI Stack
  • One console covers vulnerable AI infrastructure, risky MCP servers, malicious agent skills, workflow agents, and weak model guardrails, so security teams stop juggling five separate scanners.
  • Findings Come With Evidence
  • Each report lists a score, risk level, vulnerable file and lines, technical analysis, attack path, and remediation advice, and a reviewer agent re-verifies findings before they are shown.
  • Keep Scan Data In-House
  • Targets, uploaded code, and reports stay on your own server, and the scanning agents call only the OpenAI-compatible model endpoint you configure, so you decide what leaves the machine.
  • Extend Rules Without Code
  • Plugin management adds or edits YAML fingerprints, vulnerability entries, jailbreak datasets, and MCP detection plugins in the browser, and an update button pulls the latest rules from GitHub.

Features

  • AI Infra Scan
  • Fingerprints 100+ AI frameworks such as Ollama, ComfyUI, vLLM, and Gradio from bodies, headers, titles, and icons, then matches versions against more than 2,000 CVE-backed vulnerability rules.
  • MCP And Skill Scan
  • Three-stage agents (information collection, code audit, vulnerability review) inspect MCP servers and agent skills from a zip, GitHub URL, or SSE endpoint for tool poisoning and data theft.
  • Agent Scan For Workflows
  • Connects to Dify, Coze, or any HTTP chat endpoint, sends adversarial prompts, and reports OWASP Top 10 for agentic apps risks such as system prompt disclosure and SSRF.
  • Jailbreak Evaluation
  • Runs curated datasets with encoding and behavioral attack methods against OpenAI-compatible models, then reports a security score, risk rate, and exportable bad cases for guardrail tuning.
  • Browser Plugin Management
  • Edit fingerprint rules, vulnerability databases, evaluation datasets, and MCP plugins in the web UI, and pull updated rules from GitHub with the Update Data button.
  • Docker Compose Deployment
  • The aig-server image on port 8088 and the aig-agent image start together with docker compose on 4 GB of RAM; the UI has no login, so restrict access.

Apps Similar to AI-Infra-Guard