sj-k8s

A CLI tool for managing K8s clusters via KubeSphere API. Designed for environments where you only have KubeSphere web console access (no kubectl/SSH).

Features

• Auto login with KubeSphere's encrypted authentication (LDAP supported)
• Token caching — only re-authenticates when token expires
• Multi-cluster support with auto-discovery
• View pods, deployments, logs without kubectl access

Install

pip install sanjiu-k8s-cli

Configuration

Create ~/.k8s-cli:

KS_URL=https://your-kubesphere.example.com
KS_USER=your-username
KS_PASSWORD=your-password

Or use environment variables: KS_URL, KS_USER, KS_PASSWORD.

Usage

# Verify authentication
sj-k8s auth

# List clusters
sj-k8s clusters

# List pods (specify cluster with -c)
sj-k8s -c dev-local pods my-namespace

# Deployment status
sj-k8s -c dev-local status my-namespace my-deployment

# View logs
sj-k8s -c dev-local logs my-namespace my-pod -n 50

# Stream logs
sj-k8s -c dev-local logs my-namespace my-pod -f

# List deployments
sj-k8s -c dev-local deploys my-namespace

# Pod details
sj-k8s -c dev-local describe my-namespace my-pod

Claude Code Integration

This CLI is designed to be used with Claude Code as an AI-powered K8s operations assistant. A skill file is included in skill/SKILL.md — copy it to ~/.claude/skills/sj-k8s/SKILL.md to enable natural language K8s operations.

License

MIT