The Toolkit
AI Tools
AI is arriving in security from two directions, and this category covers both. First, agents that use security tools: MCP (Model Context Protocol) servers expose real capabilities — network scanners, exploitation frameworks, a whole toolbox — to an AI assistant, so you can drive an assessment through natural language while the agent runs the actual tools. Second, security testing of AI itself: as LLMs get deployed into products, they become an attack surface of their own, with prompt injection, jailbreaks, and data leakage to probe for.
| Tool | Source | What it’s for |
|---|---|---|
| hexstrike-ai | git + venv | An MCP server exposing ~150 security tools to an AI agent. |
| garak | AUR | A vulnerability scanner for large language models. |
| metasploit-mcp | official | Metasploit’s own MCP server, for driving the framework from an agent. |
All live under Security → AI Tools. metasploit-mcp (msfmcpd) comes with
Metasploit, so it appears under both AI
Tools and Exploitation Tools.