SYNOPSIS

# Requires a working kubeconfig and LANGERTHA_ANTHROPIC_API_KEY
LANGERTHA_ANTHROPIC_API_KEY=sk-ant-... perl -Ilib examples/raider-configmap-demo.pl

# With specific context/namespace
MCP_K8S_CONTEXT=my-cluster MCP_K8S_NAMESPACES=default \
  LANGERTHA_ANTHROPIC_API_KEY=sk-ant-... perl -Ilib examples/raider-configmap-demo.pl

DESCRIPTION

This script demonstrates MCP::K8s + Langertha::Raider working together: a real AI assistant uses the MCP tools to perform a safe, harmless operation on your Kubernetes cluster.

What it does:

1. Discovers RBAC permissions (calls k8s_permissions)
2. Creates a ConfigMap named mcp-k8s-demo with test data
3. Reads the ConfigMap back to verify it was created
4. Updates it with new data using k8s_apply
5. Deletes the ConfigMap to clean up

The entire lifecycle is driven by the AI — it decides which tools to call based on the instructions. Tool calls and timing are printed live.

Requirements:

  • LANGERTHA_ANTHROPIC_API_KEY environment variable set

  • A kubeconfig with access to at least one namespace (or MCP_K8S_TOKEN)

  • create, get, patch, delete permissions on configmaps

Safety: Only creates a single ConfigMap with test data. Cleans up after itself. No pods, deployments, or anything compute-related.

1 POD Error

The following errors were encountered while parsing the POD:

Around line 36:

Non-ASCII character seen before =encoding in '—'. Assuming UTF-8