Agent Wiz is a Python CLI tool that extracts, visualizes, and performs automated threat assessments on complex AI agent workflows across multiple orchestration frameworks.
A CLI tool for threat modeling and visualizing AI agents built using popular frameworks like LangGraph, AutoGen, CrewAI, and more.
Agent Wiz is designed for developers, researchers, and security teams who need to analyze and secure complex LLM-based agent workflows. It helps users gain visibility into agent interactions, map tool and data flows, and generate actionable security reports using established threat modeling methodologies.
Agent Wiz relies on static AST parsing and supports a variety of popular AI orchestration frameworks, making it versatile but dependent on code structure. Users should ensure their workflows are compatible with supported frameworks for accurate extraction. The tool currently integrates the MAESTRO threat modeling framework, which is specialized for AI agent security assessments.
Ensure Python is installed (compatible versions indicated on PyPI)
Install the package via pip: pip install repello-agent-wiz
Verify installation by running the CLI toolagent-wiz extract <path_to_code>
Extracts agentic workflows from the specified codebase using AST parsing.
agent-wiz visualize <extracted_workflow.json>
Generates an interactive graph visualization of agent, tool, and data flow relationships.
agent-wiz assess <extracted_workflow.json>
Performs an automated threat assessment on the extracted workflow using the MAESTRO framework.
agent-wiz export --format json <workflow_data>
Exports the workflow data and assessment results in clean JSON format for further analysis.