A.I.G

What it is

A.I.G is a self-hosted security platform from Tencent Zhuque Lab that checks AI systems for risks. It scans AI infrastructure components against a library of known CVEs, audits MCP servers and agent skills for malicious or unsafe behavior, tests agents, and runs jailbreak evaluations against LLMs, including multi-turn attacks. Security teams deploy it with Docker and use its web interface, API, or separate CLI scanners to review the AI tools and models they run.