Agentic Radar
by SPLX AI
Open-source CLI security scanner that identifies risks in AI agentic workflows
Visit Product
109 upvotes
4,049 views
About
Agentic Radar is an open-source command-line security scanner built specifically for auditing AI agentic systems and workflows. As organizations deploy increasingly autonomous AI agents with access to tools, databases, and external APIs, new categories of security risks emerge that traditional security scanners are not equipped to detect. Agentic Radar addresses this gap by analyzing AI agent code and configurations for agent-specific vulnerabilities.
The scanner checks for issues including prompt injection vulnerabilities (where malicious content in tool outputs could hijack agent behavior), privilege escalation risks (where agents have more permissions than necessary), data exfiltration pathways, and unsafe tool configurations. It analyzes popular frameworks including LangChain, LlamaIndex, AutoGPT, and CrewAI.
Agentic Radar is designed to integrate into CI/CD pipelines, enabling teams to catch security issues before deployment. Its reports provide actionable remediation guidance for each finding, helping developers understand not just what is wrong but how to fix it without breaking functionality.
The scanner checks for issues including prompt injection vulnerabilities (where malicious content in tool outputs could hijack agent behavior), privilege escalation risks (where agents have more permissions than necessary), data exfiltration pathways, and unsafe tool configurations. It analyzes popular frameworks including LangChain, LlamaIndex, AutoGPT, and CrewAI.
Agentic Radar is designed to integrate into CI/CD pipelines, enabling teams to catch security issues before deployment. Its reports provide actionable remediation guidance for each finding, helping developers understand not just what is wrong but how to fix it without breaking functionality.
Product Features
- CLI security scanner for AI agent codebases
- Prompt injection vulnerability detection
- Privilege escalation risk identification
- Tool configuration security analysis
- LangChain, LlamaIndex, CrewAI, AutoGPT support
- CI/CD integration for automated scanning
- SARIF output format for SIEM integration
- Remediation guidance for each finding
- Custom rule creation for organizational policies
- Open-source with community-contributed checks
- Prompt injection vulnerability detection
- Privilege escalation risk identification
- Tool configuration security analysis
- LangChain, LlamaIndex, CrewAI, AutoGPT support
- CI/CD integration for automated scanning
- SARIF output format for SIEM integration
- Remediation guidance for each finding
- Custom rule creation for organizational policies
- Open-source with community-contributed checks
About the Publisher
Agentic Radar was developed by SPLX AI, a security company focused on the emerging field of AI agent security. The team recognized that as AI agents move from demos to production deployments with real-world access to sensitive systems and data, security scanning must evolve beyond traditional SAST/DAST tools. SPLX AI's open-source approach to Agentic Radar reflects a belief that agentic AI security should be a shared community effort.