Claude Skills by VectorSpaceLab

"Reason about OpenQASM 3 lexical and syntactic conformance, parser

"Write, review, translate, and repair OpenQASM 3.0/3.1 programs,

"Use openqasm3 from Python to parse and print OpenQASM 3, construct

"Use OpenRLHF for Ray/vLLM/DeepSpeed RLHF workflows, including

"Prepare and validate OpenRLHF datasets for SFT, reward modeling,

"Install, diagnose, and operate OpenRLHF runtime utilities,

"Plan and adapt OpenRLHF PPO/Ray/vLLM reinforcement-learning runs,

"Build and review OpenRLHF supervised/preference training plans for

"Route OpenSquilla installation, gateway, provider and

"Operate OpenSquilla messaging channels, pairings, certification,

"Day-to-day OpenSquilla CLI automation for chat, agent runs,

"Configure OpenSquilla providers, models, SquillaRouter modes, web

"Get OpenSquilla installed, onboarded, and running through a ready

"Use this OpenSquilla sub-skill for skill catalog management,

"Router for terminal chat presentation, OpenTUI host behavior,

"Route Optax tasks across optimizer construction, loss and schedule

"Use Optax projections, tree utilities, assignment, second-order

"Compose Optax optimizers, gradient transformations, wrappers, and

"Select Optax losses, schedules, microbatching helpers, and

"Use OptiLLM as an OpenAI-compatible optimizing inference proxy

"Use OptiLLM local HuggingFace and MLX inference, LoRA adapters,

"Choose, compose, tune, and verify OptiLLM inference-time

"Configure and troubleshoot OptiLLM plugins, tool integrations,

"Operate OptiLLM as an OpenAI-compatible proxy server, including

"Routes OptiMate workflows for Speedster inference optimization,

"Guides ChatLLaMA dataset preparation, config validation,

"Guides Forward-Forward training workflows, model-type selection,

"Guides NebullVM data/device handling, backend selection, compiler

"Guides OpenAlphaTensor training, configuration, checkpointing, and

"Guides Speedster inference optimization workflows,

"Use Hugging Face Optimum for model export routing, accelerated

"Use and debug Optimum's base CLI, exporter task mappings, backend

"Operate Optimum Torch FX graph transformations and version-gated

"Configure, save, load, and troubleshoot Optimum GPT-QModel

"Operate Optimum utility APIs for dummy inputs, normalized configs,

"Use this skill for Optuna hyperparameter optimization workflows:

"Advanced Optuna support APIs for terminator stopping, search-space

"Analyze completed Optuna studies with parameter importances, trial

"Use this sub-skill for Optuna artifact upload/download workflows,

"Use this sub-skill for Optuna command-line workflows and

"Core Optuna study and trial workflows: create/load/copy/delete

"Choose, configure, customize, and debug Optuna search and pruning

"Guide future agents through Opyrator's callable wrapping, FastAPI

"Help future agents create, launch, inspect, and troubleshoot the

"Launch and reason about Opyrator's Streamlit UI, schema-driven

"Wrap Python callables with Opyrator, validate the Pydantic

"Route Orange3 data mining, modeling, visualization, and

Load, inspect, clean, reshape, validate, and save Orange data.

"Routes Orange3 exploratory plots, distance matrices, projections,

"Fit, predict with, score, and route supervised Orange3