/* SCANNER PAGE CSS - DARK BACKGROUND VERSION */ /* Replace your entire

🚀 NOW AVAILABLE - OPEN SOURCE

AgentDiscover Scanner

Find AI Agents Everywhere: Code, Network, Kubernetes

Multi-layer detection tool for comprehensive AI agent visibility across your entire infrastructure

The AI Agent Visibility Gap

90% of enterprises deploy AI agents, but only 10% have visibility into them.

Shadow AI Running Unchecked

Unmanaged AI agents in codebases, local dev environments, and production clusters operate outside governance frameworks.

No Multi-Layer Detection

Traditional tools only scan code OR monitor networks OR check clusters - not all three. You need comprehensive coverage.

Blind Spots Everywhere

Agents hiding in dependencies, running on developer laptops, or operating in production without your knowledge.

AgentDiscover Scanner Fixes It

Three Detection Layers in One Tool:

Static Code Analysis: Scans Python & JavaScript codebases to find AI frameworks (AutoGen, CrewAI, LangChain) and Shadow AI
Network Traffic Monitoring: Detects active agents by monitoring API calls to OpenAI, Anthropic, Google AI, Cohere, AWS Bedrock
Kubernetes Runtime Detection: Real-time eBPF monitoring via Cilium Tetragon with full pod/container attribution
Correlation Engine: Matches code findings with runtime behavior to classify CONFIRMED, UNKNOWN, ZOMBIE, or GHOST agents
Works Anywhere: Local repos, developer machines, CI/CD pipelines, production Kubernetes clusters

Three Ways to Use AgentDiscover Scanner

1. Scan Codebases

Use case: Security audits, compliance checks, finding Shadow AI

agent-discover-scanner scan /path/to/repo

Detects: AI frameworks, ungoverned LLM clients, risky dependencies

Perfect for: CI/CD integration, pre-commit hooks, security reviews

2. Monitor Network Traffic

Use case: Local dev machines, finding active agents in real-time

agent-discover-scanner monitor --duration 60

Detects: Live API connections to LLM providers, vector databases

Perfect for: Developer workstations, testing environments

3. Watch Kubernetes Clusters

Use case: Production visibility, runtime detection with full attribution

agent-discover-scanner monitor-k8s

Detects: Agents with pod name, container, workload, binary path

Perfect for: Production clusters, continuous monitoring, SRE teams

💡 Pro Tip: Use All Three Together

Start by scanning your codebase to find all AI frameworks. Then monitor your network to see which agents are actually running. Finally, deploy to Kubernetes clusters for production-grade visibility with full attribution. The correlation engine automatically links code findings with runtime behavior.

Get Started in 60 Seconds

Install:

# Using pipx (recommended) pipx install agent-discover-scanner # Using pip pip install agent-discover-scanner

Scan Your Code:

agent-discover-scanner scan /path/to/repo

Monitor Kubernetes:

agent-discover-scanner monitor-k8s

Example Detection:

🚨 AI Agent Detected! production/trading-bot -> OpenAI Pod: trading/high-frequency-trader-7d8f9 Workload: Deployment/trading-bot Binary: /usr/bin/python3

Enterprise-Grade Detection

Multi-Language Support

Scans Python and JavaScript/TypeScript codebases with AST-based pattern matching.

Framework Detection

Detects AutoGen, CrewAI, LangChain, LangGraph, and custom AI implementations.

Shadow AI Detection

Identifies unmanaged LLM clients bypassing governance frameworks.

Real-Time K8s Monitoring

eBPF-based detection with Cilium Tetragon integration for zero-overhead monitoring.

CI/CD Integration

SARIF output format for GitHub Security, GitLab, and other DevSecOps tools.

Full Network Coverage

Detects connections to OpenAI, Anthropic, Google AI, Cohere, Azure OpenAI, AWS Bedrock, and vector databases.

Open Source Foundation, Enterprise Power

AgentDiscover Scanner is the open-source detection layer of the DefendAI platform.

✓ Available Today (Open Source)

AgentDiscover Scanner - Find and catalog AI agents

→ Coming Q1 2026

Full DefendAI Platform with agent registry, governance, policy enforcement, and compliance reporting

Request Platform Demo

Built for Security Teams

Security Audits

Find all AI agents across your organization's codebases in minutes.

Compliance Enforcement

Detect ungoverned LLM usage that violates AI governance policies.

Production Monitoring

Real-time visibility into which AI agents are running in Kubernetes clusters.

Shadow AI Discovery

Identify AI processes running without approval or oversight.

Architecture

Detection Strategy:

Static Analysis: AST-based pattern matching for Python and JavaScript
Network Fingerprinting: Passive monitoring of LLM API connections
eBPF Monitoring: Kernel-level visibility in Kubernetes via Cilium Tetragon
Correlation: Matches code findings with runtime behavior

Requirements:

Python 3.12+
For Kubernetes monitoring: Cilium Tetragon installed, kubectl configured

Output Formats:

Terminal (rich formatted tables)
SARIF (for CI/CD integration)
JSON/JSONL (for automation)

Open Source & Community-Driven

Contribute

We welcome contributions! Check out our Contributing Guide.

Report Issues

Found a bug or have a feature request? Open an issue.

Stay Updated

Star the repo on GitHub to get notified of new releases.

Ready to Secure Your AI Agents?

Get started with the open-source scanner or request a demo of the full platform.

Try Scanner Now Request Platform Demo