AI Agent Hub
Back to plugins
🧰

dsh-kubectl

Web Tools Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install WODE25500/dsh-kubectl

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install WODE25500/dsh-kubectl in your terminal to add this plugin; the full source is at https://github.com/WODE25500/dsh-kubectl

About this plugin

Day-to-day Kubernetes operations—querying resources, reading logs, troubleshooting, deploying—have long been a blind spot in the dsh agent ecosystem. dsh-kubectl fills that gap by letting the agent invoke kubectl directly and feed the structured -o json output to the model, far more reliable than parsing plain-text tables.

On the safety side, read-only operations (get, describe, logs, status) return structured JSON with zero friction, while write operations (exec, apply, delete) and port forwarding require explicit confirmation or a persistent notice to prevent accidental changes. Configuration keys let you pin a default context, namespace, and timeout, making it well suited to multi-cluster, multi-namespace environments.

Ideal for ops engineers and platform developers who already manage Kubernetes through dsh and want their agent to cover the full query-diagnose-deploy loop. All you need is kubectl installed and a valid kubeconfig in place.

Use Cases

  • Let the agent query pod status and logs to pinpoint cluster issues quickly
  • Deploy or roll back via YAML manifests through the agent
  • Port-forward and have the agent debug in-cluster services directly

Best For

  • Ops engineers managing Kubernetes clusters with dsh
  • Platform developers involving agents in CI/CD pipelines
  • DevOps practitioners who want to drive kubectl via natural language