---

name: agents description: Manage and develop AI agents using kubani-dev CLI. Use for checking agent health, versions, deployment status, running tests, and evaluations.

AI Agents Management

Manage AI agents using the kubani-dev CLI and cluster tools.

Quick Commands

# List all agents with status
kubani-dev agents list

# Run agent locally with hot-reload
kubani-dev run k8s-monitor --hot-reload

# Run agent tests
kubani-dev test k8s-monitor

# Run evaluation suite
kubani-dev eval k8s-monitor

# View execution traces
kubani-dev trace k8s-monitor

# Start observability dashboard
kubani-dev dashboard

Arguments

• agent-name: Optional specific agent name for detailed info

Instructions

List All Agents

cd /home/al/git/kubani
echo "=== AI Agents ==="
echo ""

for earthfile in agents/*/Earthfile; do
    agent_dir=$(dirname "$earthfile")
    agent_name=$(basename "$agent_dir")
    [ "$agent_name" = "core" ] && continue

    # Get version from pyproject.toml
    version=$(grep '^version = ' "$agent_dir/pyproject.toml" | sed 's/version = "\(.*\)"/\1/')

    # Get deployed image
    deployed=$(KUBECONFIG=/home/al/.kube/config kubectl get deploy $agent_name -n ai-agents -o jsonpath='{.spec.template.spec.containers[0].image}' 2>/dev/null || echo "not deployed")

    # Get pod status
    status=$(KUBECONFIG=/home/al/.kube/config kubectl get pods -n ai-agents -l app.kubernetes.io/name=$agent_name -o jsonpath='{.items[0].status.phase}' 2>/dev/null || echo "unknown")

    echo "$agent_name"
    echo "  Source version: $version"
    echo "  Deployed image: $deployed"
    echo "  Pod status: $status"
    echo ""
done

Development Workflow

Use kubani-dev for agent development:

# Initialize configuration (one-time)
kubani-dev init

# Run agent with hot-reload
kubani-dev run k8s-monitor --hot-reload

# Run with mock services (for offline development)
kubani-dev run k8s-monitor --mock-mcp --mock-redis

# Run tests
kubani-dev test k8s-monitor --coverage

# Run evaluation suite
kubani-dev eval k8s-monitor

# Run specific evaluation layer
kubani-dev eval k8s-monitor --layer llm

Detailed Agent Info

For a specific agent, show detailed information:

AGENT_NAME="k8s-monitor"

# Pod details
KUBECONFIG=/home/al/.kube/config kubectl get pods -n ai-agents -l app.kubernetes.io/name=$AGENT_NAME -o wide

# Recent logs
KUBECONFIG=/home/al/.kube/config kubectl logs -n ai-agents -l app.kubernetes.io/name=$AGENT_NAME --tail=20

# Recent deployment history
git log --oneline -5 gitops/apps/ai-agents/$AGENT_NAME/deployment.yaml

# View traces
kubani-dev trace $AGENT_NAME --last 10

# View metrics
kubani-dev metrics $AGENT_NAME

Build and Deploy

# Build agent container
kubani-dev build k8s-monitor

# Deploy to cluster
kubani-dev deploy k8s-monitor

# Rollback deployment
kubani-dev deploy k8s-monitor --rollback

Create New Agent

# Create from default template
kubani-dev new my-agent

# Create with federated template
kubani-dev new my-agent --template federated

Core Library

The core-agents library provides shared functionality:

version=$(grep '^version = ' agents/core/pyproject.toml | sed 's/version = "\(.*\)"/\1/')
echo "core-agents (library): v$version"

# Key modules:
# - factory.py: AgentFactory, GraphFactory, DI container
# - context/: Context engineering (todo, errors, compression)
# - workflows/: Strands Graph workflow support
# - plugins/: Dynamic MCP plugin architecture
# - learning/: Continuous learning framework
# - memory/: Hierarchical memory with promotion/forgetting

Architecture

agents/
├── core/                     # Shared library
│   └── src/core_agents/
│       ├── factory.py        # AgentFactory, GraphFactory
│       ├── context/          # Context engineering
│       ├── workflows/        # Strands Graph support
│       ├── plugins/          # MCP plugin architecture
│       ├── learning/         # Continuous learning
│       └── memory/           # Hierarchical memory
├── k8s-monitor/              # Kubernetes monitoring
│   └── federated/            # Sentinel, Healer, Explorer, Triage Graph
└── news-monitor/             # News monitoring
    └── shared_agents.py      # Singleton pattern