Kubernetes Manager
A .NET MCP server that lets AI assistants manage Kubernetes clusters through natural language - list pods, scale deployments, restart services, and fetch logs without writing a single kubectl command
Ask AI about Kubernetes Manager
Powered by Claude Β· Grounded in docs
I know everything about Kubernetes Manager. Ask me about installation, configuration, usage, or troubleshooting.
0/500
Reviews
Documentation
MCP Kubernetes Manager
A .NET MCP server that lets AI assistants manage Kubernetes clusters through natural language. Instead of writing kubectl commands, you just ask Claude what you want and it talks to your cluster directly.
What you can do
Ask Claude things like:
- "List all pods in my cluster"
- "Show me a summary of the idp-platform namespace"
- "Scale the idp-platform deployment to 3 replicas"
- "Restart the idp-platform deployment"
- "Is my deployment healthy?"
- "Show me the last 50 log lines from that pod"
- "What namespaces do I have?"
Claude calls the right tool, talks to Kubernetes, and returns real data from your cluster. No kubectl, no terminal, no manual commands.
Tools
| Tool | What it does |
|---|---|
| ListPods | Lists all pods across all namespaces or a specific one |
| GetPodLogs | Fetches recent log output from any pod |
| ListDeployments | Shows all deployments with replica status |
| ScaleDeployment | Scales a deployment to a specified replica count |
| RestartDeployment | Triggers a rolling restart on a deployment |
| GetDeploymentStatus | Returns detailed rollout status and conditions |
| ListNamespaces | Lists all namespaces in the cluster |
| GetNamespaceSummary | Shows pods, deployments and services in one view |
Stack
| Layer | Technology |
|---|---|
| Framework | .NET 9 |
| MCP SDK | ModelContextProtocol 0.2.0 |
| Kubernetes client | KubernetesClient |
| Logging | Serilog |
| Transport | stdio (standard input/output) |
Running locally
Prerequisites: .NET 9 SDK, kubectl configured, Docker Desktop with Kubernetes enabled
git clone https://github.com/aftabkh4n/mcp-kubernetes-manager
cd mcp-kubernetes-manager
dotnet build
dotnet run --project src/Mcp.KubernetesManager
You should see: [INF] Connected to Kubernetes cluster: https://127.0.0.1:xxxxx [INF] MCP Server ready. Waiting for connections... [INF] Available tools: ListPods, GetPodLogs, ListDeployments...
Connecting to Claude Desktop
Open your Claude Desktop config file at:
Windows: %APPDATA%\Claude\claude_desktop_config.json
Add this:
{
"mcpServers": {
"kubernetes-manager": {
"command": "C:\\Program Files\\dotnet\\dotnet.exe",
"args": [
"run",
"--project",
"C:\\path\\to\\mcp-kubernetes-manager\\src\\Mcp.KubernetesManager"
]
}
}
}
Restart Claude Desktop. Go to Settings and you should see kubernetes-manager listed under Local MCP servers.
Open a new chat and ask: "List all pods in my Kubernetes cluster"
Safety
ScaleDeployment has a hard limit of 10 replicas to prevent accidental over-scaling. All tools are read-safe except ScaleDeployment and RestartDeployment which make changes to the cluster.
Project structure
McpKubernetes/ βββ src/ βββ Mcp.KubernetesManager/ βββ Tools/ β βββ PodTools.cs # Pod listing and log retrieval. β βββ DeploymentTools.cs # Scaling, restarting, status checks. β βββ NamespaceTools.cs # Namespace listing and summaries. βββ Program.cs # Server startup and DI registration.
CI/CD
Release notes are automatically generated by AI when a PR merges to main.
Automated release notes via Claude AI.
Release notes are generated automatically on every merge.
Automated Changelog
Every merged PR automatically generates a changelog entry via Claude AI.
Tools
- ListPods, GetPodLogs, ListDeployments, ScaleDeployment, RestartDeployment
