agentic-radar: Open Source AI Security & Privacy

What is agentic-radar?

Agentic Radar is an open-source command-line security scanner that analyzes LLM agentic workflows and produces a shareable HTML report covering their structure, tools, MCP servers, and known vulnerabilities, built for developers, researchers, and security professionals.

What it is

Agentic Radar analyzes and assesses agentic systems for security and operational insights. It is written in Python, distributed under the Apache-2.0 licence, and installs as the agentic-radar command. It reads workflow definitions from five supported agentic frameworks — langgraph, crewai, n8n, openai-agents, and autogen — and inspects them to explain how a system functions and where it may be exposed.

The concrete thing it replaces is ad hoc manual review. Rather than a developer reading framework-specific definitions to reconstruct which tools an agent can call and which MCP servers it reaches, the scanner emits a report: a workflow graph, a list of external and custom tools, a list of MCP servers used by the system's agents, and a vulnerability mapping table connecting identified tools to known vulnerabilities. Those findings map to two published frameworks, the OWASP Top 10 for LLM Applications and OWASP Agentic AI – Threats and Mitigations.

Key capabilities

agentic-radar scan accepts a framework argument — one of langgraph, crewai, n8n, openai-agents, or autogen — reads a source folder, and writes an HTML report, for example agentic-radar scan langgraph -i path/to/langgraph/example/folder -o report.html.
agentic-radar test exercises agents in a workflow for vulnerabilities; it currently supports openai-agents and requires OPENAI_API_KEY set as an environment variable.
Workflow Visualization renders a graph of the agentic system's workflow.
Tool Identification lists every external and custom tool the system uses.
MCP Server Detection lists all MCP servers used by the system's agents.
Vulnerability Mapping produces a table linking identified tools to known vulnerabilities.
Agentic Prompt Hardening rewrites detected system prompts into structured form and displays them in the report.

Who uses it and how

Development teams running agentic workflows through CI/CD pipelines, matching the project's devsecops and agentic-workflow topics and its CI/CD Workflow section.
Security professionals and AI red teams auditing LLM deployments, matching the ai-red-teaming and llm-security topics.
Researchers who need to understand how an agentic system actually functions before drawing conclusions about it.
Teams running MCP servers, since detection of those servers is a first-class output rather than a side effect.
Users who only need the workflow picture can skip the scanner and use the hosted Agentic Visualizer at agentic-visualizer.splx.ai.

Getting started

Install with pip install agentic-radar, then confirm with agentic-radar --version. Framework extras install separately: pip install "agentic-radar[crewai]" and pip install "agentic-radar[openai-agents]".

How it compares

The facts list no paid products that this project replaces, and they name no directly comparable scanner. On the evidence provided, it stands alone in this registry.

When to use it — and when not to

The CrewAI extras pull in crewai-tools, which is only supported on Python 3.10 and above; on older interpreters tool descriptions will be less detailed or missing entirely. The test command is limited to the openai-agents framework and depends on an external OpenAI API key, so vulnerability testing carries a cost the scan command does not. Teams working outside the five supported frameworks should not pick it up, because there is nothing for it to parse.

project readme (upstream, from github) — read inline

A Security Scanner for your agentic workflows!

View Demo · Documentation · Report Bug · Request Feature

Table of Contents

Description
Agentic Visualizer
Getting Started
- Prerequisites
- Installation
Advanced Installation
- CrewAI Installation
- OpenAI Agents Installation
Usage
Advanced Features
Roadmap
Blogs and Tutorials
Community
Frequently Asked Questions
Contributing
Code Of Conduct
License

Description 📝

The Agentic Radar is designed to analyze and assess agentic systems for security and operational insights. It helps developers, researchers, and security professionals understand how agentic systems function and identify potential vulnerabilities.

It allows users to create a security report for agentic systems, including:

Workflow Visualization - a graph of the agentic system's workflow✅
Tool Identification - a list of all external and custom tools utilized by the system✅
MCP Server Detection - a list of all MCP servers used by system's agents✅
Vulnerability Mapping - a table connecting identified tools to known vulnerabilities, providing a security overview✅

The comprehensive HTML report summarizes all findings and allows for easy reviewing and sharing.

View Full Report Example Here

Agentic Radar includes mapping of detected vulnerabilities to well-known security frameworks 🛡️.

Agentic Visualizer 🎆

If you only care about visualization, try out the Agentic Visualizer.

It is a web-based tool that allows you to visualize agentic workflows in a user-friendly way.

Getting Started 🚀

Prerequisites

There are none! Just make sure you have Python (pip) installed on your machine.

Installation

pip install agentic-radar

# Check that it is installed
agentic-radar --version

Some features require extra installations, depending on the targeted agentic framework. See more below.

Advanced Installation

CrewAI Installation

CrewAI extras are needed when using one of the following features in combination with CrewAI:

Agentic Radar Test
Descriptions for predefined tools

You can install Agentic Radar with extra CrewAI dependencies by running:

pip install "agentic-radar[crewai]"

[!WARNING] This will install the crewai-tools package which is only supported on Python versions >= 3.10 and If you are using a different python version, the tool descriptions will be less detailed or entirely missing.

OpenAI Agents Installation

OpenAI Agents extras are needed when using one of the following features in combination with OpenAI Agents:

Agentic Radar Test

You can install Agentic Radar with extra OpenAI Agents dependencies by running:

pip install "agentic-radar[openai-agents]"

Usage

Agentic Radar now supports two main commands:

1. `scan`

Scan code for agentic workflows and generate a report.

agentic-radar scan [OPTIONS] FRAMEWORK:{langgraph|crewai|n8n|openai-agents|autogen}

Example:

agentic-radar scan langgraph -i path/to/langgraph/example/folder -o report.html

2. `test`

Test agents in an agentic workflow for various vulnerabilities. Requires OPENAI_API_KEY set as environment variable.

agentic-radar test [OPTIONS] FRAMEWORK:{openai-agents} ENTRYPOINT_SCRIPT_WITH_ARGS

Example:

agentic-radar test openai-agents "path/to/openai-agents/example.py"

See more about this feature here.

Advanced Features ✨

Agentic Prompt Hardening

Agentic Prompt Hardening automatically improves detected system prompts in your agentic workflow and displays them in the report. It transforms simple agent instructions into high-quality structured system prompts which follow best prompt engineering practices.

[!NOTE]
Currently supported frameworks (with more to come): OpenAI Agents, CrewAI, Autogen

It is quite straightforward to use:

Set your OPENAI_API_KEY environment variable by running export OPENAI_API_KEY=.
Run Agentic Radar with the --harden-prompts flag, for example:

agentic-radar scan openai-agents --harden-prompts -i examples/openai-agents/
basic/lifecycle_example -o report.html

Inspect hardened system prompts in the generated report:

🔍 Test for Vulnerabilities in Agentic Workflows

Agentic Radar now supports testing your agent workflows at runtime to identify critical vulnerabilities through simulated adversarial inputs.

This includes automated testing for:

Prompt Injection
PII Leakage
Harmful Content Generation
Fake News Generation

Currently supported for:

OpenAI Agents ✅ (more frameworks coming soon)

🛠 How It Works

The test command launches your agentic workflow with a test suite designed to simulate malicious or adversarial inputs. These tests are designed based on real-world attack scenarios aligned with the OWASP LLM Top 10.

[!NOTE]
This feature requ

Frequently asked questions

Is agentic-radar free to use?

agentic-radar is open source under the Apache-2.0 licence. There is no licence fee and no seat count — you can self-host it or, where the project offers one, pay a vendor for a managed version instead.

What does agentic-radar do?

A security scanner for your LLM agentic workflows

What is agentic-radar written in?

agentic-radar is primarily written in Python. Its source is publicly available at https://github.com/splx-ai/agentic-radar, and it has 1,054 GitHub stars.