This release adds 1 notable feature for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Topics
+5 more
Affected surfaces
Summary
AI summaryFixed in‑cluster Kubernetes config loading and now raises a clear ProviderError when neither kubeconfig nor in‑cluster config is available.
Full changelog
Fix: In-Cluster Kubernetes Config
Resolves the HTTPConnectionPool(host='localhost', port=80): Max retries exceeded error when the server is deployed inside a Kubernetes cluster (#61).
What Changed
- Fixed config loading —
_patched_load_kube_configno longer short-circuits calls with explicit arguments (config_file, context, client_configuration), which was preventing the kubernetes client from loading the correct API server address - In-cluster fallback — The kubeconfig provider now automatically falls back to in-cluster config when no kubeconfig file exists in the container
- Actionable errors — Raises
ProviderErrorwith clear remediation steps when neither kubeconfig nor in-cluster config is available, instead of silently falling back tolocalhost:80 - Deployment manifest — Added
MCP_K8S_PROVIDER=in-clusterandKUBERNETES_SERVICE_PORT=443env vars
Deployment
For in-cluster deployments, ensure these env vars are set:
env:
- name: MCP_K8S_PROVIDER
value: "in-cluster"
- name: KUBERNETES_SERVICE_HOST
value: "kubernetes.default.svc"
- name: KUBERNETES_SERVICE_PORT
value: "443"
Reference manifests: deploy/kubernetes/
Install
pip install kubectl-mcp-server==1.23.0
# or
npx [email protected]
# or
docker pull rohitghumare64/kubectl-mcp-server:1.23.0
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About rohitg00/kubectl-mcp-server
A Model Context Protocol (MCP) server for Kubernetes that enables AI assistants like Claude, Cursor, and others to interact with Kubernetes clusters through natural language.
Related context
Beta — feedback welcome: [email protected]