Claude Skills by VectorSpaceLab

"Use when working with the LiteLLM AI Gateway proxy: starting the

"Use this when configuring, validating, or debugging LiteLLM Router

"Use for direct LiteLLM Python SDK work: chat/text completions,

"Use LitGPT to run local LLM inference, chat, finetuning,

"Download, validate, convert, classify, and merge LitGPT

"Evaluate LitGPT checkpoints with LM Evaluation Harness and serve

"Use LitGPT for local text generation, chat, and Python API

"Plan and validate LitGPT finetuning, LoRA/QLoRA/full/adapter

"Serve custom AI APIs with LitServe, including general LitAPI and

"Expose LitServe LitAPI endpoints as Model Context Protocol tools

"Build and debug LitServe OpenAI-compatible chat completions and

"Serve models and custom pipelines with LitServe LitAPI and

"Use for LlamaFactory repository workflows: training configs,

"Use this sub-skill when working on LlamaFactory dataset_info

"Use this sub-skill for LlamaFactory inference, ChatModel usage,

"Use for LlamaFactory model/tokenizer/processor loading, model

"Use for LlamaFactory v0 training runs, config files, CLI

"Use for LlamaFactory experimental v1 flows selected with USE_V1=1,

"Operate LLaMA Factory's LlamaBoard Web UI, web chat,

"Build, customize, troubleshoot, and maintain LlamaIndex Python

"Build LlamaIndex agents and workflows with tools, memory, chat

"Customize LlamaIndex Settings, prompts, callbacks,

"Build and query LlamaIndex indexes, persist and reload storage,

"Load local files or raw text into LlamaIndex Documents, split them

"Choose and verify LlamaIndex integration packages, namespace

"Maintain the LlamaIndex monorepo safely: inspect package layout

"Router for LlamaGen autoregressive image-generation workflows,

"Router for LlamaGen class-conditional ImageNet generation,

"Router for ImageNet code extraction, T5 feature extraction, and

"Router for LlamaGen caption-conditioned training, sampling, prompt

"Router for VQ, VQGAN, Stable Diffusion VAE, and Consistency

"Guides LLaVA vision-language model inference, serving, training,

"Guides LLaVA image chat, local inference, CLI prompts, model

"Guides LLaVA custom VQA evaluation, benchmark inference, chunked

"Guides LLaVA custom training data, pretraining, full fine-tuning,

"Use MosaicML LLM Foundry for LLM data preparation, Composer

"Prepare LLM Foundry pretraining, fine-tuning, raw-text,

"Configure and run LLM Foundry offline and in-training ICL

"Use LLM Foundry inference helpers for Hugging Face generation,

"Inspect and use LLM Foundry package APIs, registries, MPT/HF model

"Create, inspect, adapt, and troubleshoot LLM Foundry pretraining

"Route bilingual concept lookup, diagram-family navigation,

Safely inspect, preview, and maintain the bilingual diagram asset tree.

"Route concept lookup, bilingual terminology, diagram-family

"Route requests for LLM4Decompile training, direct decompilation,

"Run direct decompilation inference and benchmark scoring for

"Extract pseudo-code with Ghidra and refine it with the

"Run the SK²Decompile two-phase skeleton/skin pipeline, RL helpers,

Train, fine-tune, or prepare datasets for LLM4Decompile models.

"Use EleutherAI LM Evaluation Harness for language-model evaluation