Agentic Radar
VerifiedSecurity scanner that maps and audits risks in agentic AI workflows.
What is Agentic Radar?
Agentic Radar is an open-source scanner built to inspect agentic applications for security gaps and operational details. It generates a single HTML report that includes a visual workflow graph, lists of external and custom tools, detected MCP servers, and a table that ties each tool to known vulnerabilities.
Users run the scanner against existing agent codebases. The resulting report highlights potential issues and maps them to established security frameworks, making it easier to prioritize fixes. Additional capabilities include prompt hardening suggestions and integration into CI/CD pipelines.
The project targets developers, security engineers, and researchers who build or maintain multi-agent systems and need a repeatable way to surface risks without manual code review.
Capabilities
What you can build with Agentic Radar
Pre-deployment security review
Run scans on agent workflows to surface tool-related vulnerabilities and receive an HTML report that can be shared with security teams.
Workflow visualization
Generate a clear graph of how agents, tools, and MCP servers interact so teams can understand data flows and dependencies at a glance.
CI/CD integration
Add the scanner to automated pipelines to catch new vulnerabilities whenever agent code changes are merged.
Install Agentic Radar
pip install agentic-radarpip install agentic-radar
# Check that it is installed
agentic-radar --version- 1Install the package with pip install agentic-radar.
- 2Install framework-specific extras for CrewAI or OpenAI Agents if needed.
- 3Point the scanner at your agent project directory.
- 4Review the generated HTML report for workflow graphs and vulnerability tables.
- 5Add the scan command to your CI configuration for ongoing checks.
Agentic Radar: pros & cons
Pros
- +Produces a single, shareable HTML report with visualizations and mappings.
- +Detects both external tools and MCP servers used by agents.
- +Links findings to recognized security frameworks like OWASP.
- +Open-source with PyPI distribution and community channels.
Cons
- –Requires separate installation steps for each supported agent framework.
- –Report quality depends on how thoroughly the scanner can parse the codebase.
- –Limited to the frameworks explicitly supported at install time.
Frequently asked questions
It supports CrewAI and OpenAI Agents through optional installation extras.
User reviews
Verified reviews from the community shape this listing's rating.
Loading reviews…
Similar agents
Other automation options worth comparing.
