Skill UI

A powerful skill covering geospatial science, including GIS, remote sensing, spatial statistics, and machine learning for Earth observation. It supports processing of various data types (raster, vector, point cloud) from sources like Sentinel and Landsat, and integrates modern cloud-native workflows (STAC, COG). Ideal for environmental monitoring, urban planning, and advanced spatial modeling tasks.

A comprehensive guide for advanced geospatial data processing using GeoPandas. This skill targets professional workflows involving planar vector data, emphasizing rigorous data provenance, coordinate reference system (CRS) management, and complex spatial operations. It details necessary safety contracts, correctness gates, and best practices for handling sensitive geographic information in reproducible Python environments.

Gtars is a robust toolkit providing native Rust implementations and Python bindings for advanced genomic interval modeling. It supports core bioinformatic tasks including set algebra, coverage calculation, tokenization, and fragment processing. The library emphasizes data integrity by enforcing strict genomic data contracts, ensuring correct handling of coordinates, assemblies, and reference sequences across complex computational pipelines.

A comprehensive framework for transforming raw observations into rigorous, evidence-bounded, and testable research plans. It guides users through establishing research questions (PICO/PECO), formulating candidate hypotheses and rival explanations, defining causal estimands, and creating preregistration-ready analysis plans, all while maintaining strict ethical and scientific guardrails.

LaminDB is an open-source, lineage-native lakehouse designed for managing complex biological datasets and computational workflows. It makes diverse scientific data (like scRNA-seq, AnnData) queryable, traceable, and compliant with FAIR principles. Core capabilities include robust data lineage tracking, schema validation, ontology-backed annotation (Bionty), and unified querying across multiple cloud/local storage systems. It is essential for bioinformatics research and building reproducible data infrastructure.

Open Notebook is a self-hosted, open-source alternative to Google NotebookLM, designed for advanced research and document analysis. It allows users to ingest diverse content (PDFs, videos, web pages) and organize them into private notebooks. Key features include AI-powered summarization, context-aware chatting, vector search, and multi-speaker podcast generation, all while ensuring complete data sovereignty by running entirely on your own infrastructure and supporting multiple AI providers.

polars-bio is a high-performance Python library built on Polars for genomic interval operations and bioinformatics file I/O. It provides a familiar DataFrame-centric API for interval arithmetic (overlap, nearest, merge, complement, etc.) and supports reading/writing major bioinformatics formats like BED, VCF, BAM, and GFF. Key features include streaming, cloud-native I/O, and SQL query support, making it a fast alternative for large-scale genomic data processing.

This guide provides comprehensive, version-aware documentation for using PufferLib in reinforcement learning pipelines. It details best practices for adapting Gymnasium/PettingZoo environments, managing different PufferLib versions (3.0.0 vs 4.0 source), and executing safe, reproducible training and evaluation routines. Key focus areas include environment contract validation, sandboxing native builds, and maintaining data provenance.

RDKit is a comprehensive cheminformatics library providing robust Python APIs for molecular structure manipulation and analysis. It supports tasks ranging from molecular I/O (SMILES, SDF) and calculating physical descriptors (MW, LogP) to advanced substructure searching, 3D coordinate generation, and chemical reaction modeling. It is an essential toolkit for drug discovery, computational chemistry, and academic cheminformatics research.

scikit-bio is a comprehensive Python library designed for advanced bioinformatics and biological data analysis. It provides robust tools for sequence manipulation (DNA, RNA, Protein), performing sophisticated sequence alignments (global, local, multiple), constructing and analyzing phylogenetic trees, and conducting complex ecological statistics. Key features include calculating diversity metrics (alpha/beta), running ordination analyses (PCoA), and handling diverse biological file formats (FASTA, Newick, BIOM). Ideal for researchers in microbiology, genomics, and ecology.

SHAP (SHapley Additive exPlanations) is a powerful technique used to interpret and audit machine learning model predictions. It calculates feature attributions (SHAP values) to explain how input features map to model outputs. This skill guides users through selecting appropriate explainers, handling multi-output models, validating feature attributions, and generating visualizations for both local and global model understanding.

The Hugging Face Transformers library is a comprehensive toolkit providing access to thousands of pre-trained models across diverse domains including Natural Language Processing (NLP), Computer Vision, Audio, and Multimodal tasks. It enables developers to easily load models, perform high-level inference (e.g., text generation, classification, QA), and fine-tune models on custom datasets using streamlined APIs like Pipelines and the Trainer class. Ideal for building advanced, state-of-the-art AI applications.