Does this skill support containerized environments?

Yes, it includes built-in support for Singularity (Apptainer) containers, allowing you to wrap job commands with specific images and bind paths.

Which workload managers does this skill support?

This skill is specifically designed for clusters using the SLURM workload manager in conjunction with the MDClaw CLI tools.

How does it handle resource management on shared clusters?

You can set resource policies that restrict allowed partitions, maximum GPU counts, and memory limits to ensure jobs stay within your project's allocation.

Can it help me if my job hits a wall-time limit?

Yes, the skill can analyze the job's state, identify the timeout, and generate a restart script that continues the simulation from the last saved checkpoint.

HPC SLURM Job Manager

Name: HPC SLURM Job Manager
Author: matsunagalab

bymatsunagalab

•

クラウドインフラストラクチャ

Automates SLURM-based high-performance computing workflows, including job submission, cluster inspection, and error recovery for scientific simulations.

The HPC Run skill enables Claude to manage the full lifecycle of High-Performance Computing (HPC) tasks using the SLURM workload manager and MDClaw CLI. It simplifies complex scientific workflows by discovering cluster resources, enforcing resource policies, and generating optimized job scripts for Molecular Dynamics (MD) or AI-driven structure prediction like Boltz-2. Designed for researchers and engineers, it automates the tedious aspects of cluster management, including real-time job monitoring, log-based error diagnosis, and seamless checkpoint-based restarts for interrupted simulations.

主な機能

01Automated SLURM job submission and real-time status monitoring

02Domain-specific script generation for MD simulations and protein folding

03Intelligent cluster discovery and resource policy enforcement

04Advanced error recovery with automated log analysis and diagnosis

051 GitHub stars

06Support for Singularity container integration and environment module loading

ユースケース

01Submitting and monitoring long-running Molecular Dynamics simulations with optimized HMR parameters

02Automating protein structure prediction pipelines using GPU-accelerated HPC partitions

03Diagnosing failed batch jobs and performing restarts from specific checkpoint files

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add matsunagalab/mdclaw hpc-run

For use in Claude.ai and ChatGPT