Agent Skills

AKO4ALL — Agentic Kernel Optimization

ako4all

Automated loop that profiles, iterates, benchmarks and commits GPU kernel optimizations across CUDA/Triton/TileLang/Python/C++ to achieve measurable speedups.

HPC Python Patterns

claude-skill-registry

Guidance and patterns for Python parallelism and GPU/CPU performance: threading vs multiprocessing vs asyncio, CUDA streams, PyTorch DDP, and benchmarking.

MoE Expert-Parallel Overlap

megatron-bridge

Guides enabling and validating MoE expert-parallel communication overlap in Megatron-Bridge to hide dispatch/combine latency and improve throughput.

Graphsignal GPU Profiler

graphsignal-profiler

Set up and manage GPU profiling, tracing, and monitoring for AI inference workloads using vLLM, SGLang, and PyTorch.

GPU Kernel Bottleneck Analyzer

pjt222

Systematically identify if a GPU kernel is compute-bound, memory-bound, or latency-bound.

PyTorch Pretrained Weight Loader

asb-skill-collections

Load pretrained PyTorch model checkpoints to rank candidate molecules or score spectra in metabolomics without retraining.