skills.rest
zhuyingqin
zhuyingqinCOMMUNITY
Explore all skills by this author →
GitHub

qzcli

Manage GPU compute jobs on the Qizhi platform via kubectl-style qzcli commands.

1|1|Updated May 19, 2026
Search Tags:#job submission#distributed training#mcp integration#gpu jobs#qzcli#kubectl-style cli#hpc orchestration
One-click install
npx skills add https://github.com/zhuyingqin/ARIS-WEB --skill qzcli-zhuyingqin
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: qzcli
Source: https://github.com/zhuyingqin/ARIS-WEB/tree/main/crates/runtime/assets/skills/qzcli
Command: npx skills add https://github.com/zhuyingqin/ARIS-WEB --skill qzcli-zhuyingqin

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) components.

What problem does it solve?

qzcli removes the friction of logging into the Qizhi (启智) platform and manually tracking GPU compute resources and job lifecycles by providing a single, kubectl/docker-style command workflow for discovery, submission, and control.

Core Features & Use Cases

  • Workspace & compute-group discovery: Lists and refreshes cached workspaces, compute groups, and available nodes so you can target the right resources.
  • Job submission & parameterization: Creates GPU compute jobs interactively (TUI selection) or non-interactively (IDs or names) including command, image, instances, priority, spec, and dry-run.
  • Operational controls: Stops jobs, inspects status/details, and watches running jobs; also supports batch submission via a JSON config matrix for repeatable experiments.

Quick Start

Ask the assistant to “Use qzcli to list available compute resources and then submit a test GPU job using my workspace and compute group, returning the new job id.”

Frequently Asked Questions about qzcli

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I submit and manage distributed training jobs on Qizhi without using the web UI?▼

You can manage Qizhi GPU jobs using a kubectl-style CLI to discover compute resources, submit training workloads, and control the job execution lifecycle entirely from the command line.

What is the best way to run batch GPU job submissions for repeatable experiments on HPC clusters?▼

Batch GPU job submission is best handled using a JSON config matrix to define repeatable experiments, allowing you to submit multiple distributed training jobs non-interactively with varying parameters.

How do I discover available compute groups and node specs before submitting a GPU job?▼

You can list and refresh cached workspaces, compute groups, and available nodes using CLI discovery commands, ensuring you target the right resources before submitting your GPU job.

Can I use kubectl-style commands to monitor and stop running distributed training jobs?▼

Yes, you can inspect job status, watch running jobs in real-time, and stop execution using dedicated kubectl-style operational controls designed for the Qizhi platform.

Does GPU job submission on Qizhi support interactive parameter selection?▼

Yes, GPU compute jobs can be created interactively using TUI selection or non-interactively using specific IDs and names for compute groups, specs, and images.

Why do I need to refresh the resource cache before listing workspaces for HPC orchestration?▼

Refreshing the resource cache ensures you have the most current availability data for workspaces and compute nodes, which is required after logging in before you can accurately discover resources and submit jobs.

Related Skills

Explore More Software Engineering Skills→
Shallow-WShallow-W

qzcli

Manage GPU compute job lifecycle on the Qizhi platform via kubectl-style CLI.

Community
Intermediate
yunqiqiliangyunqiqiliang
8

clickzetta-monitoring

Monitor ClickZetta Lakehouse job status and performance via SHOW JOBS and information_schema views.

Community
Advanced
duypham2801duypham2801

qzcli

Manage GPU compute jobs on the Qizhi platform via a kubectl-style CLI.

Community
Intermediate
Mang30Mang30

qzcli

Manage GPU compute jobs on the Qizhi platform via CLI.

Community
Intermediate