Introducing Agent-Wiz : Threat Modeling for AI Agents

🚀 Release: Agent-Wiz v0.4.0 – Your CLI Companion for AI Agent Visualization & Threat Modeling

We're thrilled to announce the initial release of Agent-Wiz : a powerful command-line tool designed to help developers, researchers, and security practitioners visualize and threat model AI agents with ease.

🎯 What is Agent-Wiz?

Agent-Wiz is a CLI tool that empowers you to:

🧠 Map and visualize the architecture of AI agents, their decision flows, and interactions.

🔍 Identify vulnerabilities and model potential threats across different stages of the agent lifecycle.

⚡ Streamline your AI security audits by integrating threat modeling directly into your development workflow.

✨ Features

✅ Interactive CLI interface with step-by-step prompts

🗺️ Agent architecture diagram generation (DOT, PNG, SVG)

🛡️ Built-in MAESTRO framework based threat modeling support

📦 Exportable reports (Markdown, JSON)

🧪 Compatible with ALL popular agent frameworks (LangGraph, Autogen, CrewAI, Swarm, Pydantic and more!)

What's Changed

• [FEAT + BUG] added crew ai multi agent-tool example from EchoSynth by @justary27 in #10
• [BUG] Remove unkown edge mapping in langgraph by @Aaditya-G in #11
• [FEAT] Adds openai example for an ecommerce agent by @Aaditya-G in #12

New Contributors

• @Aaditya-G made their first contribution in #11

Full Changelog: https://github.com/Repello-AI/Agent-Wiz/commits/0.4.0