Automate experiment execution on the TACC Vista HPC cluster.
npx skills add https://github.com/plantcad/genecad --skill tacc
Please help me install this Agent Skill. Skill: tacc Source: https://github.com/plantcad/genecad/tree/main/.claude/skills/tacc Command: npx skills add https://github.com/plantcad/genecad --skill tacc
Manages experiment execution on the TACC Vista HPC cluster. Use when the user says "run on TACC", "run the pipeline", "allocate a GPU node", "check SLURM queue", "sync code to TACC", "download results from TACC", "cancel that job", or when any task requires GPU/CPU compute that cannot run locally. Do NOT use for local-only tasks like editing code, running tests, or plotting.
Set up SSH access, install the command wrapper, verify the environment, and run your TACC experiments using the provided workflow.
High-intent search queries and answers about installing and using this skill.
To automate running experiments on the TACC Vista HPC cluster, this workflow validates SSH access, syncs code, allocates GPU nodes, and manages SLURM queue execution end-to-end. It wraps commands to ensure proper environment propagation on compute nodes.
Yes, you can sync code and download results from TACC automatically. The workflow manages the entire lifecycle of an HPC experiment, including pushing local repositories to the cluster and retrieving generated outputs once SLURM jobs complete.
Allocating GPU nodes and monitoring SLURM queues are core capabilities. The workflow automates resource allocation on TACC Vista and tracks job status, allowing you to manage multi-node execution using srun and log outputs directly to SCRATCH.
Running pipelines on TACC Vista requires SSH access setup and environment verification. The workflow first validates your SSH connection and checks repository existence on the cluster before executing any compute tasks or allocating resources.
You should not use this HPC experiment workflow for local-only tasks like editing code, running tests, or plotting. It is specifically designed for tasks requiring heavy GPU or CPU compute that cannot be executed locally on your machine.
czh-ee-2023Deploy and execute machine learning experiments across local and remote environments.
InternScienceParse Gaussian gjf input files and convert them to XYZ format.
yo-stevenConvert single-node PyTorch distributed launches into multi-node Slurm sbatch jobs.
xjtulycContainerize research environments with Docker, Compose, and Singularity for reproducible workflows.
