gpu-resource-optimizer

Gpu Resource Optimizer - Auto-activating skill for ML Deployment. Triggers on: gpu resource optimizer, gpu resource optimizer Part of the ML Deployment skill category.

1.05x

Quality

Does it follow best practices?

Impact

99%

1.05x

Average score across 3 eval scenarios

Securityby

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./planned-skills/generated/08-ml-deployment/gpu-resource-optimizer/SKILL.md

Evaluation results

100%

GPU Memory Optimization for a PyTorch Inference Service

Production-ready GPU memory optimization

Criteria

Without context

With context

Dynamic batching

100%

GPU memory limit config

100%

torch.cuda.empty_cache or memory cleanup

100%

Half-precision inference

100%

No-grad context

100%

Batch size in config

100%

Error handling for OOM

50%

100%

Step-by-step deployment instructions

100%

Hardware-specific guidance

100%

Production config separation

100%

Timeout parameter

100%

Without context: $0.4098 · 2m 26s · 14 turns · 15 in / 8,863 out tokens

With context: $0.6724 · 3m 26s · 27 turns · 319 in / 11,921 out tokens

100%

Multi-GPU Training Pipeline for a CV Team

Multi-GPU MLOps pipeline configuration

Criteria

Without context

With context

Distributed training setup

100%

Per-job GPU allocation

100%

Memory limit per GPU

100%

Checkpointing

100%

Job queue configuration

100%

Hardware spec in config

100%

Step-by-step setup instructions

100%

Environment variable GPU control

100%

No-grad in validation

100%

Timeout or max runtime config

100%

Separate config from code

100%

Without context: $0.4778 · 2m 54s · 16 turns · 16 in / 11,267 out tokens

With context: $0.8707 · 4m 3s · 29 turns · 30 in / 16,762 out tokens

97%

GPU Health Monitoring for a Production ML Serving Cluster

GPU monitoring and production alerting

Criteria

Without context

With context

nvidia-smi integration

100%

Zero-utilization detection

100%

Memory leak / stale process detection

80%

70%

Structured log output

75%

100%

Configurable polling interval

100%

Threshold-based alerting

100%

Alert severity levels

100%

Configurable thresholds

100%

Step-by-step runbook

100%

Remediation steps

100%

Output path config

100%

Without context: $0.7309 · 3m 44s · 22 turns · 23 in / 15,056 out tokens

With context: $0.7904 · 4m 19s · 25 turns · 182 in / 16,110 out tokens

Repository: jeremylongshore/claude-code-plugins-plus-skills
Commit: f17dd51

Evaluated: 18 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

GPU Memory Optimization for a PyTorch Inference Service Multi-GPU Training Pipeline for a CV Team GPU Health Monitoring for a Production ML Serving Cluster

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.