Skill UI

Statsmodels is Python's premier library for rigorous statistical modeling. It provides comprehensive tools for estimation, inference, and diagnostics across a wide range of statistical methods. Use this skill for advanced analysis, including linear regression (OLS, WLS), generalized linear models (Logistic, Poisson), and complex time series forecasting (ARIMA). Ideal for econometric studies and producing publication-ready results.

OpenSquilla-compatible adapter that generates browser-playable AI audio such as voiceovers, narration, music, and sound effects through OpenRouter audio models. It streams OpenAI gpt-audio output, writes WAV assets, and reports structured AUDIO_READY manifests for webpage integration.

This skill audits OpenMed NER or de-identification models to detect performance disparities across sensitive demographic subgroups (e.g., sex, age, race). It calculates metrics like per-group recall and leakage disparity, helping users identify instances of 'under-protection' before deployment, ensuring equitable AI governance and compliance.

This skill automates the creation of comprehensive Model Cards for OpenMed clinical NER or de-identification models. It gathers data from various OpenMed evaluation reports (e.g., `GateReport`, `fairness_report`, `error_report`) to structure critical governance information, including intended use, quantitative performance metrics, subgroup analysis, limitations, and required medical-device disclaimers. It ensures documentation is reproducible and compliant for clinical AI deployment.

This tool benchmarks the recall of PII detection models, crucial for ensuring privacy compliance before deployment. It uses synthetic fixtures and enforces a recall floor by measuring label-aware exact-span and grapheme recall against predefined gold standards. It is essential for auditing models and preventing critical privacy leaks.

This skill provides an in-depth, entity-level scorecard for clinical or biomedical Named Entity Recognition (NER) models. It calculates precision, recall, and F1 scores (both strict and partial match modes) using a user-supplied gold corpus. Key features include per-label error breakdown, a confusion matrix, and detailed False Negative/False Positive analysis, essential for debugging model performance in highly sensitive medical domains.

This skill converts unstructured documents (PDF, EPUB, DOCX, etc.) into actionable, structured agent skills. Instead of providing simple summaries, it deep-dives to extract core frameworks, mental models, key principles, techniques, and anti-patterns from educational material. Use this when you need to transform academic knowledge into a reusable, machine-readable toolkit for agents like Copilot CLI, Amp, or Claude Code.

This skill scaffolds a gold-standard annotation project tailored for evaluating clinical NER and de-identification models (OpenMed). It guides users through defining label schemas, writing rigorous guidelines, configuring annotation tools (like BRAT/Label Studio), generating synthetic clinical text, and creating disciplined, non-leaking train/dev/test splits. It ensures that all committed data are synthetic fixtures for reproducible evaluation.

This skill helps developers programmatically discover, filter, and inspect the entire OpenMed model registry. It allows users to find the correct model key for specific clinical tasks (e.g., diseases, oncology, PII detection), filter by language, size, or task, and review model metadata before loading. It pairs with `loading-openmed-models`.

This tool de-identifies sensitive free-text columns (e.g., names, dates, identifiers) in local structured datasets (CSV, JSONL, Parquet) using advanced OpenMed models. It ensures strict privacy by generating a completely redacted dataset and a separate PHI-free aggregate summary, allowing for secure data sharing and analysis without compromising source privacy or logging sensitive values.

This skill de-identifies sensitive PII from clinical notes across numerous languages, including Spanish, German, French, and Arabic. It automatically selects language-specific PII models, national ID regex patterns, and generates locale-aware fake surrogates. Designed for global healthcare datasets, it runs entirely on-device, ensuring maximum privacy and linguistic accuracy for diverse international clinical records.

The Diagram Generator converts complex, unstructured inputs—such as natural language, code snippets, schemas, and notes—into professional, editable visual diagrams. It supports diverse standards including Mermaid, PlantUML, and Graphviz, enabling the creation of flowcharts, sequence diagrams, UML models, ER diagrams, architecture maps, and more. Outputs are designed to be reviewable, version-controllable, and highly precise.