Skill UI

A meta-skill dedicated to creating high-quality, publication-ready scientific figures. It guides users on best practices for journal submission, including multi-panel layouts, precise error bar inclusion, colorblind-safe palette usage, and adherence to specific journal styles (Nature, Science). It orchestrates libraries like matplotlib, seaborn, and plotly to ensure optimal resolution and format export (PDF, EPS, TIFF).

TorchDrug is a comprehensive PyTorch-based ML toolbox designed for molecular and protein science. It leverages Graph Neural Networks (GNNs) to tackle complex tasks including molecular property prediction, protein structure modeling, drug-target binding prediction, knowledge graph reasoning, and de novo molecular generation. It provides robust capabilities for accelerating drug discovery research.

This skill provides comprehensive access to 78 public REST APIs, covering biomedical science, chemistry, materials science, physics, and economics. Users can query vast datasets for genes, proteins, compounds, clinical trials, patents, and environmental indicators. It is an indispensable tool for deep academic research, complex data analysis, and cross-disciplinary scientific investigation.

This skill provides comprehensive tools and methodologies for working with the Brain Imaging Data Structure (BIDS) standard. It is essential for organizing, validating, querying, and converting complex multi-modal neuroscience and biomedical datasets (including MRI, EEG, MEG, PET, etc.). Use this when setting up BIDS-compliant directory structures, ensuring data integrity before publication, or converting raw DICOM scans into standardized formats.

This skill converts complex, narrative technical documents, research excerpts, or natural language problem descriptions into a highly structured Mathematical Problem Specification (MPS). It systematically extracts variables, constraints, objectives, and uncertainty parameters. By adhering to a strict Zero-Inference Protocol, it ensures all extracted elements are strictly grounded in the source text, making it invaluable for formalizing real-world optimization, data science, and engineering problems.