Kubernetes

Kubernetes Skill

Comprehensive Kubernetes cluster management through native kubectl execution. This skill provides 18+ kubectl tools for managing your clusters.

When to Use

• Managing Kubernetes resources (pods, deployments, services, configmaps, secrets)
• Viewing pod logs and debugging containers
• Scaling deployments up or down
• Managing cluster configuration and contexts
• Performing rollout operations (restart, undo, status)
• Node management (cordon, uncordon, drain, taint)
• Creating and deleting resources

Requirements

• kubectl must be installed and in PATH
• Valid kubeconfig (default: ~/.kube/config or KUBECONFIG env var)
• Appropriate cluster access permissions

Tools Provided

get

Get Kubernetes resources (pods, services, deployments, nodes, etc.)

Parameters:

• resource (required): Resource type (pods, services, deployments, nodes, namespaces, configmaps, secrets, ingress, pv, pvc, jobs, cronjobs, daemonsets, statefulsets, replicasets, events, endpoints, all)
• name (optional): Specific resource name
• namespace (optional): Kubernetes namespace (default: default, use 'all' for all namespaces)
• selector (optional): Label selector (e.g., app=nginx)
• output (optional): Output format (wide, yaml, json, name)

Example:

skill run kubernetes get resource=pods namespace=kube-system
skill run kubernetes get resource=deployments output=wide
skill run kubernetes get resource=nodes

describe

Show detailed information about a resource.

Parameters:

• resource (required): Resource type (pod, service, deployment, node, etc.)
• name (required): Resource name
• namespace (optional): Kubernetes namespace

Example:

skill run kubernetes describe resource=pod name=nginx-xxxxx namespace=default

logs

Get logs from a pod.

Parameters:

• pod (required): Pod name
• namespace (optional): Kubernetes namespace
• container (optional): Container name (for multi-container pods)
• tail (optional): Number of lines to show from end
• follow (optional): Stream logs in real-time
• previous (optional): Show logs from previous container instance

Example:

skill run kubernetes logs pod=nginx-xxxxx namespace=default tail=100

exec

Execute a command in a container.

Parameters:

• pod (required): Pod name
• namespace (optional): Kubernetes namespace
• container (optional): Container name
• command (required): Command to execute

Example:

skill run kubernetes exec pod=nginx-xxxxx command="ls -la /var/log"

apply

Apply a configuration from YAML/JSON content.

Parameters:

• content (required): YAML or JSON content to apply
• namespace (optional): Kubernetes namespace
• dry_run (optional): Run in dry-run mode

Example:

skill run kubernetes apply content="apiVersion: v1\nkind: ConfigMap\nmetadata:\n  name: test-config\ndata:\n  key: value"

delete

Delete Kubernetes resources.

Parameters:

• resource (required): Resource type
• name (required): Resource name
• namespace (optional): Kubernetes namespace
• force (optional): Force deletion
• grace_period (optional): Grace period in seconds

Example:

skill run kubernetes delete resource=pod name=nginx-xxxxx namespace=default
skill run kubernetes delete resource=namespace name=test-namespace

scale

Scale a deployment, replicaset, or statefulset.

Parameters:

• resource (required): Resource type (deployment, replicaset, statefulset)
• name (required): Resource name
• namespace (optional): Kubernetes namespace
• replicas (required): Desired replica count

Example:

skill run kubernetes scale resource=deployment name=nginx namespace=default replicas=3

rollout

Manage rollouts for deployments.

Parameters:

• resource (required): Resource type (deployment, daemonset, statefulset)
• name (required): Resource name
• namespace (optional): Kubernetes namespace
• action (required): Rollout action (status, history, undo, restart, pause, resume)

Example:

skill run kubernetes rollout resource=deployment name=nginx action=status
skill run kubernetes rollout resource=deployment name=nginx action=restart

top

Display resource usage (CPU/memory).

Parameters:

• resource (required): Resource type (pods, nodes)
• namespace (optional): Kubernetes namespace
• containers (optional): Show container-level metrics

Example:

skill run kubernetes top resource=pods namespace=default
skill run kubernetes top resource=nodes

cluster-info

Display cluster information.

Example:

skill run kubernetes cluster-info

config

Manage kubeconfig (view, current-context, get-contexts, use-context).

Parameters:

• action (required): Config action (view, current-context, get-contexts, use-context)
• context (optional): Context name (for use-context)

Example:

skill run kubernetes config action=view
skill run kubernetes config action=current-context
skill run kubernetes config action=use-context context=minikube

create

Create resources (namespace, secret, configmap, deployment, service).

Parameters:

• resource (required): Resource type
• name (required): Resource name
• namespace (optional): Kubernetes namespace
• from_literal (optional): Key=value pairs for configmap/secret
• image (optional): Container image (for deployment)
• port (optional): Port number
• type (optional): Secret type or Service type

Example:

skill run kubernetes create resource=namespace name=test-namespace
skill run kubernetes create resource=deployment name=nginx namespace=test image=nginx:latest port=80
skill run kubernetes create resource=configmap name=myconfig namespace=default from_literal="key1=value1,key2=value2"

label

Add or update labels on resources.

Parameters:

• resource (required): Resource type
• name (required): Resource name
• namespace (optional): Kubernetes namespace
• labels (required): Comma-separated key=value pairs

Example:

skill run kubernetes label resource=deployment name=nginx namespace=default labels="env=prod,tier=frontend"

annotate

Add or update annotations on resources.

Parameters:

• resource (required): Resource type
• name (required): Resource name
• namespace (optional): Kubernetes namespace
• annotations (required): Comma-separated key=value pairs

Example:

skill run kubernetes annotate resource=deployment name=nginx namespace=default annotations="description=Web server"

cordon

Mark a node as unschedulable.

Parameters:

• node (required): Node name

Example:

skill run kubernetes cordon node=worker-node-1

uncordon

Mark a node as schedulable.

Parameters:

• node (required): Node name

Example:

skill run kubernetes uncordon node=worker-node-1

drain

Drain a node for maintenance.

Parameters:

• node (required): Node name
• ignore_daemonsets (optional): Ignore DaemonSet pods
• delete_emptydir_data (optional): Delete pods using emptyDir
• force (optional): Force drain

Example:

skill run kubernetes drain node=worker-node-1 ignore_daemonsets=true

taint

Add a taint to a node.

Parameters:

• node (required): Node name
• taint (required): Taint in key=value:effect format

Example:

skill run kubernetes taint node=worker-node-1 taint="dedicated=gpu:NoSchedule"

raw

Execute any kubectl command directly.

Parameters:

• args (required): Raw kubectl arguments

Example:

skill run kubernetes raw args="version --client"
skill run kubernetes raw args="api-resources"
skill run kubernetes raw args="get pods -A -o wide"

Configuration

This skill uses your existing kubectl configuration. No additional configuration is required.

To use a specific kubeconfig:

export KUBECONFIG=/path/to/kubeconfig

Or configure in .skill-engine.toml:

[skills.kubernetes.instances.prod]
config.kubeconfig = "${KUBECONFIG:-~/.kube/config}"

Security Notes

• This skill executes real kubectl commands against your cluster
• Commands are validated against an allowlist (kubectl, helm, etc.)
• Ensure your kubeconfig has appropriate permissions
• Be cautious with delete and drain operations

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: kubiyabot
• Source: kubiyabot/skill
• License: Apache-2.0

Install and usage instructions live in the source repository linked above.