Validate the inventory file: it parses and passes structural validation, every referenced environment variable (${VAR}) is set, and every SSH key / kubeconfig file it references exists and is readable. Re-reads the inventory file from disk on every call, so it reflects the file's current state — including drift since the server started — not just what's currently loaded in memory. Read-only, takes no arguments.

Fetch a Cisco device's running configuration ("show running-config"), as-is. Telnet-connected devices get "terminal length 0" sent first automatically; SSH-connected devices do not (SSH runs one command per ephemeral session, so the setting wouldn't carry over) — if an SSH-connected device paginates output, only the first page is returned unless "terminal length 0" is already configured on the device.

Fetch a Cisco device's running configuration and diff it against the last snapshot taken by this tool for the same device — flags unexpected config drift. The first call for a device has nothing to diff against and just establishes the baseline.

List a Cisco device's interfaces with IP address, admin/link status, and line protocol state ("show ip interface brief").

List a Cisco device's hardware components — chassis, modules, PSUs — with PID/VID/serial number ("show inventory").

Fetch a Cisco device's buffered syslog messages ("show logging"), optionally capped to the most recent limit lines.

Get a Cisco device's software version, hostname, and uptime ("show version").

Report aggregate CPU utilization on a Linux server, sampled over a 1-second window, with a warning/critical severity and recommendation.

Report per-filesystem disk usage on a Linux server, with a warning/critical severity and recommendation for filesystems running low on space.

List local Docker images on a Docker host.

Fetch the most recent combined stdout/stderr log lines from a container on a Docker host.

List containers on a Docker host (running only, unless all is set).

Restart a container on a Docker host. This is a dangerous, service-impacting action: it requires confirm: true, and returns without acting if confirmation is missing. Set dry_run: true to see the exact command that would run without needing confirm at all.

Report a point-in-time CPU/memory/network/disk usage snapshot for containers on a Docker host.

List systemd services currently in the failed state on a Linux server.

List currently firing or resolved alert instances from a Grafana instance's alerting API.

List annotations on a Grafana instance, optionally scoped to a time range and/or tags.

Search dashboards on a Grafana instance, optionally filtered by title text or tag.

Run a raw query against one datasource on a Grafana instance. The response shape is datasource-specific (Prometheus, Loki, SQL, ...); this is a passthrough, not a normalized result.

Report the most recent systemd journal entries at error priority or higher on a Linux server.

Report the running kernel release (uname -r) of a Linux server.

Describe a pod: its spec/status highlights plus recent events for it.

List recent Kubernetes events, most recent first, optionally scoped to one namespace.

List pods in a Kubernetes cluster, optionally scoped to one namespace.

Fetch log lines from a pod's container in a Kubernetes cluster.

List nodes in a Kubernetes cluster with readiness, roles, and capacity.

List servers known to the inventory, optionally filtered by tag. Never returns credentials.

Report memory usage on a Linux server, with a warning/critical severity and recommendation when memory is running low.

List every node in a Proxmox VE cluster along with its resource usage summary.

Take a snapshot of a QEMU VM or LXC container on a Proxmox node. This is a state-changing action: it requires confirm: true, and returns without acting if confirmation is missing. Snapshotting is asynchronous — the response carries a task UPID, not confirmation the snapshot has finished; check it with proxmox_tasks. Set dry_run: true to see the exact API call that would be made without needing confirm at all.

Start a QEMU VM or LXC container on a Proxmox node. This is a state-changing action: it requires confirm: true, and returns without acting if confirmation is missing. Starting is asynchronous — the response carries a task UPID, not confirmation the guest has finished starting; check it with proxmox_tasks. Set dry_run: true to see the exact API call that would be made without needing confirm at all.

Force-stop (hard power-off, not a graceful shutdown) a QEMU VM or LXC container on a Proxmox node. This is a dangerous, service-impacting action: it requires confirm: true, and returns without acting if confirmation is missing. Stopping is asynchronous — the response carries a task UPID, not confirmation the guest has finished stopping; check it with proxmox_tasks. Set dry_run: true to see the exact API call that would be made without needing confirm at all.

List recent tasks (backups, migrations, VM actions, ...) recorded on a Proxmox node, most recent first.

List every QEMU VM and LXC container on a Proxmox node.

Report whether a Linux server is pending a reboot (Debian/Ubuntu reboot-required marker, or a running kernel older than the newest installed one).

List the top processes on a Linux server, sorted by CPU usage descending.

Report how long a Linux server has been running and its 1/5/15-minute load averages.