SuperAPI AI Directory
Partner with us中文AI newsFavorites
All toolsAI agentsAgent-Wiz

Agent-Wiz

Agent-Wiz is a Python command-line tool for extracting agent workflows from mainstream AI frameworks such as LangGraph, AutoGen, and CrewAI, visualizing them, and automatically performing security threat assessments based on threat modeling methodologies such as MAESTRO. It is aimed at developers, researchers, and security teams, helping to clarify the relationships among agents, tools, and data flows in complex LLM orchestration, identify potential vulnerabilities, and generate security reports. Agent-Wiz requires an OpenAI API key to be configured in order to run analysis commands. Related framework parsers, standardized graph representations, security report generation, analysis for additional threat models such as STRIDE/PASTA/LINDDUN, and agent simulation-based threat exploration are still planned or under development.

Open website

About this tool

Purpose: Agent-Wiz is a Python-based command-line tool (CLI) for extracting agent workflows from mainstream AI agent orchestration frameworks and performing automated threat assessments. It is aimed at developers, researchers, and security teams, with the goal of providing visibility into complex LLM-based orchestration, helping visualize flows, map tool and agent interactions, and generate actionable security reports.

Main capabilities:

  1. Workflow extraction: Provides dedicated AST static parsers for supported frameworks to extract class-based or function-based agents, transitions between agents, tool call chains, and group agents (such as selector, round-robin).
  2. Visualization: Generates HTML d3-based visualizations of agent workflows; the optional open flag automatically opens the visualization in the default browser.
  3. Security analysis: Currently uses MAESTRO as the main threat modeling framework, assessing agent workflows according to the structure of Mission (system purpose and security objectives), Assets (inventory of key components such as agents, tools, and data flows), Entrypoints (attack surface and access vector mapping), Security Controls (assessment of existing defenses), Threats (identification of potential vulnerabilities and attack scenarios), Risks (calculation of impact and likelihood of security incidents), and Operations (runtime security assessment).
  4. Report generation: Generates reports based on the provided graph and threat modeling framework, such as autogen_report.md.

Usage:

  • Before running any analysis command, an OpenAI API key must first be configured, which can be set either via environment variables or a .env file (recommended).
  • Extract agent workflows: This command generates a graph JSON with a specific structure.
  • Visualize agent workflows: This command generates an HTML visualization based on d3.
  • Threat modeling analysis: This command generates a report similar to autogen_report.md.
  • You can run agent-wiz --help for more information.
  • For persistent use, you can add the relevant configuration lines to .bashrc, .zshrc, or an environment setup script.

Applicable scenarios: Suitable for developers, researchers, and security teams building agent systems with frameworks such as LangGraph, AutoGen, and CrewAI, for clarifying agent and tool relationships in complex LLM orchestration, discovering potential securityvulnerabilities, and producing security assessment reports.

Limitations and plans:

  • Currently mainly supports the MAESTRO threat modeling framework.
  • Analysis for additional threat models such as STRIDE, PASTA, and LINDDUN is still under development.
  • Planned features (in no particular order) include: building parsers for major agent frameworks, generating a standardized JSON graph representation of agent flows, security report generation, extending to STRIDE/PASTA/LINDDUN, etc., and agent simulation-based threat exploration.

Related tools

Browse category