Where do you start?

Most engineers do the same thing — kubectl get pods --all-namespaces, kubectl describe, grep through logs. An hour later you have pieces of the picture but not the full story.

There is a faster way

steered — a single binary that analyzes your cluster live

steered runs on your local machine — wherever kubectl is configured. No agents, no cloud, no setup. Just run it and it gives you the full picture.

Install it in one command:

curl -fsSL https://steered.dev/install | sudo sh

Then run: steered

What steered finds

The moment steered starts it begins analyzing your cluster live:

• CVEs — checks your running images against known vulnerabilities

• Misconfigurations — missing resource limits, privileged containers, exposed secrets

• Security issues — missing network policies, RBAC misconfigurations

• Good practice violations — what will hurt you later if not fixed

• Health score — real time percentage of cluster health

No manual cross-checking. No grepping through docs. Everything surfaced in seconds.

It analyzes. It guides. You fix.

Hit 'a' and steered opens the AI analysis view.

For every issue it finds:

• WHY — root cause in plain English

• ACTION — what to do about it

• RUN — the exact kubectl command, ready to copy

• RISK — what happens if you ignore it

Works with ollama locally, openai, or anthropic. The better the model, the sharper the analysis.

Zero setup. Your machine. Your rules.

steered follows the same kubeconfig precedence as kubectl. If kubectl works on your machine — steered works.

No cloud account. No agents on the cluster. No database. Single Go binary for Linux, Mac, and Windows.

The knowledge base is community-driven — markdown files called skills. Anyone can contribute detection rules, CVE advisories, and good practice checks.

Update its skills anytime. Keep it sharp.

Try it:

curl -fsSL https://steered.dev/install | sudo sh

Website: https://steered.dev
GitHub: https://github.com/steereddev/steered

kubernetes devops golang securities opensource