Follow these steps to run an MCP Kubernetes agent and tool in a Meshagent Room:

​
1. Install the Meshagent CLI with all features:

pip install "meshagent[all]"

​
2. Sign up and authenticate your account:

See Getting Started with the CLI.

​
3. Start an MCP server in a Meshagent Room:

meshagent service test --room=test --role=agent --image=meshagent/mcp-kubernetes:latest --env MESHAGENT_PORT=8001  --port="num=8001 path=/webhook liveness=/ type=meshagent.callable" --name=mcp-kubernetes-service
  • This command launches a Meshagent Room with the MCP Kubernetes server available inside the room.
  • Meshagent Rooms will close if they go inactive.

​
4. Start a chatbot in the room, with access to the Kubernetes toolkit:

meshagent chatbot join --room=test --agent-name=sample --name=sample --toolkit=mcp-kubernetes
  • This starts a chatbot agent in the Meshagent Room, capable of using the Kubernetes toolkit.
  • You can use multiple toolkits with the same agent.
  • A link to the room is provided in the command output—follow it to interact live with the agent and tools.

​
Tools Available

The MCP Kubernetes Server provides the following tools for remote Kubernetes management:

ToolShort Description
cleanupCleanup all managed resources
explain_resourceGet documentation for a Kubernetes resource or field
install_helm_chartInstall a Helm chart
kubectl_applyApply a Kubernetes YAML manifest from a string or file
kubectl_contextManage Kubernetes contexts - list, get, or set the current context
kubectl_createCreate Kubernetes resources using various methods (from file or using subcommands)
kubectl_deleteDelete Kubernetes resources by resource type, name, labels, or from a manifest file
kubectl_describeDescribe Kubernetes resources by resource type, name, and optionally namespace
kubectl_genericExecute any kubectl command with the provided arguments and flags
kubectl_getGet or list Kubernetes resources by resource type, name, and optionally namespace
kubectl_listList Kubernetes resources by resource type and optionally namespace
kubectl_logsGet logs from Kubernetes resources like pods, deployments, or jobs
kubectl_patchUpdate field(s) of a resource using strategic merge patch, JSON merge patch, or JSON patch
kubectl_rolloutManage the rollout of a resource (e.g., deployment, daemonset, statefulset)
kubectl_scaleScale a Kubernetes deployment
list_api_resourcesList the API resources available in the cluster
port_forwardForward a local port to a port on a Kubernetes resource
stop_port_forwardStop a port-forward process
uninstall_helm_chartUninstall a Helm release
upgrade_helm_chartUpgrade a Helm release

​
Tool Details

​
cleanup

Cleanup all managed resources

​
explain_resource

Get documentation for a Kubernetes resource or field
Parameters:

  • resource (string): Resource name or field path (e.g. ‘pods’ or ‘pods.spec.containers’)
  • apiVersion (string, optional): API version to use (e.g. ‘apps/v1’)
  • output (string, optional): Output format (plaintext or plaintext-openapiv2)
  • recursive (boolean, optional): Print the fields of fields recursively

​
install_helm_chart

Install a Helm chart
Parameters:

  • chart (string): Chart name
  • name (string): Release name
  • namespace (string): Kubernetes namespace
  • repo (string): Chart repository URL
  • values (object, optional): Chart values

​
kubectl_apply

Apply a Kubernetes YAML manifest from a string or file
Parameters:

  • dryRun (boolean, optional): Only validate the resource without applying
  • filename (string, optional): Path to a YAML file to apply
  • force (boolean, optional): Immediately remove resources and bypass graceful deletion
  • manifest (string, optional): YAML manifest to apply
  • namespace (string, optional): Namespace

​
kubectl_context

Manage Kubernetes contexts
Parameters:

  • operation (string): list/get/set
  • detailed (boolean, optional): Include detailed info
  • name (string, optional): Name of the context to set (for set)
  • output (string, optional): Output format
  • showCurrent (boolean, optional): Highlight the current context when listing

​
kubectl_create

Create Kubernetes resources
Parameters:

​
kubectl_delete

Delete Kubernetes resources
Parameters:

​
kubectl_describe

Describe Kubernetes resources
Parameters:

  • name (string): Name of the resource
  • resourceType (string): Type of resource
  • allNamespaces (boolean, optional)
  • namespace (string, optional)

​
kubectl_generic

Execute arbitrary kubectl commands
Parameters:

  • command (string)
  • Additional args and flags

​
kubectl_get

Get or list Kubernetes resources
Parameters:

  • resourceType (string)
  • allNamespaces, fieldSelector, labelSelector, name, namespace, output, sortBy (all optional)

​
kubectl_list

List Kubernetes resources
Parameters:

  • resourceType (string)
  • Optional: allNamespaces, fieldSelector, labelSelector, namespace, output

​
kubectl_logs

Get logs from Kubernetes resources
Parameters:

  • name, namespace, resourceType (all string)
  • Several optional parameters for granularity

​
kubectl_patch

Patch Kubernetes resources
Parameters:

  • name, resourceType (both string)
  • Optional: dryRun, namespace, patchData, patchFile, patchType

​
kubectl_rollout

Manage the rollout of a resource
Parameters:

  • name, resourceType, subCommand (all string)
  • Optional: namespace, revision, timeout, toRevision, watch

​
kubectl_scale

Scale a deployment
Parameters:

  • name (string)
  • replicas (number)
  • Optional: namespace, resourceType

​
list_api_resources

List resources
Parameters:

  • Optional: apiGroup, namespaced, output, verbs

​
port_forward

Forward ports
Parameters:

  • localPort (number)
  • resourceName, resourceType (string)
  • targetPort (number)
  • Optional: namespace

​
stop_port_forward

Stop a port-forward
Parameters:

  • id (string)

​
uninstall_helm_chart

Uninstall a Helm chart
Parameters:

  • name (string)
  • namespace (string)

​
upgrade_helm_chart

Upgrade a Helm release
Parameters:

  • chart, name, namespace, repo (string)
  • Optional: values (object)

​
Meshagent & MCP Resources

By wrapping the MCP server in Meshagent Rooms, Kubernetes administrators, developers, and AI agents can share secured access to robust cluster management tools and automation in a modern, collaborative environment.