Skill UI

A comprehensive tool to manage deprecated fields and update query patterns for the Fireflies.ai GraphQL API. Use this skill when integrating with Fireflies and encountering breaking changes, such as updating single-value parameters to array formats or migrating deprecated fields.

High-performance RLHF framework that combines Ray, vLLM, and ZeRO-3 to accelerate PPO/GRPO/RLOO/DPO training for 7B-70B+ models on distributed GPU clusters, providing a single-stack workflow for reward models and policy optimization.

This skill provides expert guidance on writing highly portable shell scripts that adhere strictly to POSIX standards. It teaches users how to avoid shell-specific extensions (like advanced arrays, specific conditionals, or GNU features) to ensure maximum compatibility and robustness across diverse Unix-like systems (Linux, BSD, Solaris, etc.). Ideal for critical infrastructure scripting and environments requiring extreme cross-platform reliability.

This skill provides a comprehensive pattern for optimizing bulk API interactions. It enables efficient processing of large arrays of operations by supporting both synchronous (for small batches) and asynchronous (for large batches) execution. Key features include concurrency control, robust partial failure handling, progress tracking via background workers, and dedicated validation steps, ensuring reliable data throughput for high-volume requests.

Ray Data offers scalable distributed data processing for ML workloads, streaming across CPU/GPU clusters and integrating with PyTorch, TensorFlow, and Ray Train. Use it for batch inference, ETL, and multi-modal preprocessing from a laptop up to hundreds of nodes.

Ray Train orchestrates distributed PyTorch, TensorFlow, and HuggingFace workloads from laptops to clusters, automating GPU placement, fault tolerance, checkpointing, and hyperparameter sweeps so teams can scale massive model training with minimal code changes.

A comprehensive guide for writing efficient GLSL shaders (Vertex/Fragment) used in WebGL and game engines. Learn fundamental concepts like shader structure, uniforms, varying variables, and advanced techniques such as raymarching and post-processing effects, optimizing graphics rendering performance on the GPU.

Guides training large language models with Volcano Engine's verl library, covering PPO/GRPO workflows, distributed backends, multi-turn rollouts, reward configs, critic-enabled PPO, and production deployments using Ray workers.

Dask is a Python library that scales pandas/NumPy workflows beyond RAM by parallelizing DataFrames, Arrays, Bags, and custom tasks across cores or clusters for ETL, log analysis, and distributed ML.

FlowIO reads and writes Flow Cytometry Standard (FCS) files, exposing metadata, channels, and NumPy event arrays while allowing CSV/DataFrame conversion, making it ideal for cytometry pipelines and backend preprocessing services.

PyLabRobot is a vendor-agnostic, pure Python SDK designed for autonomous and automated laboratory workflows. It provides a unified interface to control a wide array of laboratory equipment, including liquid handling robots (Hamilton, Opentrons, Tecan), plate readers, pumps, shakers, and incubators. Ideal for developing complex, multi-device, reproducible protocols and simulating experiments before physical execution.

TileDB-VCF lets you ingest, store, query, and export large genomic VCF/BCF datasets using TileDB’s sparse arrays, supporting incremental sample growth, parallel regions queries, and cloud-local interoperability for population genomics workflows.