AI assistants

An AI assistant that can run shell commands can use cv4pve-metrics-exporter in two ways: read the endpoint to answer questions with the values of this moment (which guests are running, how full a storage is, which guests have no backup job), and help you run the exporter as a service and scrape it from Prometheus.

The cv4pve-metrics-exporter skill

The repository ships a skill, skills/cv4pve-metrics-exporter/SKILL.md: a short file in the Agent Skills format that Claude Code, Codex and other assistants load when a request matches its description. It tells the assistant to:

• read the endpoint of an exporter that already runs, and start one only when none does, stopping it afterwards;
• read the endpoint once and work on that copy: every request reads the whole cluster;
• show you a change to a service, to the Prometheus configuration or to the listening address before making it.

Install it with the skills command, which needs Node.js:

npx skills add Corsinvest/cv4pve-metrics-exporter

Or copy the file into the skills folder of your assistant: .claude/skills/ for Claude Code,.agents/skills/ for Codex, in the project or under your home folder:

mkdir -p .claude/skills/cv4pve-metrics-exporter
curl -fL -o .claude/skills/cv4pve-metrics-exporter/SKILL.md https://raw.githubusercontent.com/Corsinvest/cv4pve-metrics-exporter/master/skills/cv4pve-metrics-exporter/SKILL.md

Give it a token of its own

The skill tells the assistant what to do; the token decides what it can do. The assistant can read the token and call the Proxmox VE API without cv4pve-metrics-exporter, and the skill cannot prevent that. Give it a read-only token: PVEAuditor is all the exporter needs, see Permissions. Put the connection in an options file, so the token stays out of the conversation. An assistant that only reads the endpoint of a running exporter needs no token at all.