MCP server for interacting with Kubernetes clusters via kubectl
36 tools with significant quality gaps. Most tools have basic descriptions (10-50 chars, well below the 194-char baseline) that lack context for LLM selection. Parameter descriptions are sparse or missing entirely, many parameters have no explanation of their purpose, format, or constraints. Input schemas are present but lack per-parameter descriptions. No output schemas documented. Error handling is minimal. Tool names follow verb_noun convention (good), but descriptions fail to explain WHEN to use each tool or what it returns. No tool annotations (readOnlyHint/destructiveHint). Destructive tools (delete_*, cleanup) lack confirmation patterns. READ_ONLY tools lack pagination guidance despite list_* tools accepting namespace filters.
Cleanup all managed resources
Create a new Kubernetes ConfigMap
Create a new Kubernetes CronJob
Create a new Kubernetes deployment
Create a new Kubernetes namespace
Create a new Kubernetes pod
Create a new Kubernetes service
Delete a Kubernetes CronJob
Destructive tools (delete_*, cleanup, uninstall_helm_chart) lack confirmation/dry-run patterns. No guidance for LLM to verify intent before irreversible operations.
Parameter descriptions missing or minimal across all tools. E.g., 'namespace' parameter in list_pods has no description explaining it filters by namespace. LLMs cannot infer parameter semantics from names alone.
No output schemas documented. Tools return unstructured responses; LLMs cannot plan downstream calls or extract required fields (e.g., pod_id, deployment_name) for chaining.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 39 | 2026-07-28+ | v2 |
Delete a Kubernetes deployment
Delete a Kubernetes namespace
Delete a Kubernetes pod
Describe a Kubernetes CronJob
Describe a Kubernetes deployment
Describe a Kubernetes pod
Explain a Kubernetes resource
Get the current Kubernetes context
Get Kubernetes events
Get logs from a Kubernetes job
Get logs from Kubernetes resources
Install a Helm chart
List Kubernetes API resources
List Kubernetes contexts
List Kubernetes CronJobs
List Kubernetes deployments
List Kubernetes jobs
List all namespaces
List Kubernetes nodes
List Kubernetes pods
List Kubernetes services
Scale a Kubernetes deployment
Set the current Kubernetes context
Start port forwarding to a Kubernetes resource
Stop port forwarding
Uninstall a Helm chart
Update a Kubernetes deployment
Upgrade a Helm chart
Tool descriptions are too brief (10-50 chars vs. 194-char baseline). Lack WHEN to use, prerequisites, and what is returned. E.g., 'Describe a Kubernetes pod' does not explain when to call it vs. get_logs or list_pods.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). LLMs cannot distinguish safe read-only tools from destructive ones without explicit hints.