tencent/ai-infra-guard

A full-stack AI Red Teaming platform securing AI ecosystems via Agent Scan, Skills Scan, MCP scan, AI Infra scan and LLM jailbreak evaluation.

View on GitHub ↗Jump to charts ↓Open shareable report

Summary Information

Updated 50 minutes ago
Added to GitGenius on August 20th, 2026
Created on December 25th, 2024
Open Issues & Pull Requests: 22 (+0)
Number of forks: 477
Total Stargazers: 4,843 (+27)
Total Subscribers: 38 (+0)

Repository Insights (GitGenius)

Median issue/PR response: 1.9 hours
Mean response time: 4.0 days
90th percentile: 3.6 days
Tracked items: 122

How this project is maintained

Around half of the issues opened in the past year never receive a reply. Only 14% of issues opened in the past year have been closed. Three people close 90% of everything that gets resolved.

Charts & Analytics

Fetching additional details & charts...

Issue Activity (beta)

Open issues: 12
New in 7 days: 9
Closed in 7 days: 4
Avg open age: 62 days
Stale 30+ days: 6
Stale 90+ days: 1

Recent activity

Opened in 7 days: 8
Closed in 7 days: 4
Comments in 7 days: 4
Events in 7 days: 11

Top labels

  • enhancement (34)
  • bug (17)
  • question (8)
  • discussion (7)
  • good first issue (6)
  • documentation (4)
  • invalid (2)
  • feature (1)

Detailed Description

AI-Infra-Guard is a full-stack AI red teaming platform that secures AI ecosystems through multiple scanning and evaluation capabilities.

The platform addresses the need to identify security vulnerabilities across AI infrastructure and applications. It works by integrating five distinct scanning approaches: Agent Scan for evaluating agent security, Skills Scan for assessing AI skill implementations, MCP Scan for Model Context Protocol server security, AI Infra Scan for infrastructure vulnerabilities, and Jailbreak Evaluation for testing language model robustness against prompt injection and other adversarial techniques. This multi-layered approach allows organizations to conduct comprehensive security assessments across their entire AI stack rather than examining individual components in isolation.

Teams should adopt this tool if they operate AI agents, language models, or AI infrastructure and need systematic vulnerability assessment. It suits organizations building production AI systems where security evaluation is a requirement rather than an afterthought. The platform provides both breadth across different AI system types and depth within each scanning category, making it applicable to diverse AI deployment scenarios from simple LLM applications to complex multi-agent systems.

The project's maintainers respond to issues and pull requests within hours, indicating active engagement with the user community. Work tracked in the issue system centers on enhancements, bug fixes, and user questions, reflecting a development cycle focused on expanding capabilities and addressing reported problems.