sue-nm5-slurm-status

What problem does it solve?

This Skill eliminates the need to manually run multiple Slurm commands to check NM5 cluster scheduler state, job status, pending reasons, and GPU-hour accounting, while eliminating the risk of accidental job mutations during status checks.

Core Features & Use Cases

• Read-Only Slurm Snapshots: Retrieve live queue state, running job details, pending job reasons, and 3-day job history for NM5 Slurm clusters without modifying any scheduler state.
• GPU-Hour Accounting & Quota Tracking: Calculate physical GPU-hour consumption, efficiency metrics, remaining/projected GPU usage, and configured quota allocation status for your compute accounts.
• Integrated Reporting: Generate redacted local HTML reports for detailed job inspection, or supply paste-ready fragments for the sue-nm5-monitor skill's comprehensive NM5 status report.
• Use Case: If you are running ML training jobs on NM5, use this Skill to quickly diagnose pending job blockers, track your GPU quota usage, and verify job progress without learning complex Slurm command syntax.

Quick Start

Invoke the sue-nm5-slurm-status skill to retrieve a read-only snapshot of your NM5 Slurm queue, job status, and GPU-hour accounting metrics.