Synchronize repository documentation

Co-authored-by: Cursor <cursoragent@cursor.com>

README.md

@@ -1,8 +1,8 @@
 # Scientific Agent Skills
 
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE.md)
-[![Version](https://img.shields.io/badge/Version-2.53.0-blue.svg)](pyproject.toml)
-[![Skills](https://img.shields.io/badge/Skills-148-brightgreen.svg)](#-whats-included)
+[![Version](https://img.shields.io/badge/Version-2.54.0-blue.svg)](pyproject.toml)
+[![Skills](https://img.shields.io/badge/Skills-150-brightgreen.svg)](#-whats-included)
 [![Databases](https://img.shields.io/badge/Databases-100%2B-orange.svg)](#-whats-included)
 [![Agent Skills](https://img.shields.io/badge/Standard-Agent_Skills-blueviolet.svg)](https://agentskills.io/)
 [![Security Scan](https://github.com/K-Dense-AI/scientific-agent-skills/actions/workflows/security-scan.yml/badge.svg)](https://github.com/K-Dense-AI/scientific-agent-skills/actions/workflows/security-scan.yml)
@@ -23,11 +23,11 @@
 
 > **🔔 Claude Scientific Skills is now Scientific Agent Skills.** Same skills, broader compatibility — now works with any AI agent that supports the open [Agent Skills](https://agentskills.io/) standard, not just Claude.
 
-> **New: [K-Dense BYOK](https://github.com/K-Dense-AI/k-dense-byok)** — A free, open-source AI co-scientist that runs on your desktop, powered by Scientific Agent Skills. Bring your own API keys, pick from 40+ models, and get a full research workspace with web search, file handling, 100+ scientific databases, and access to all 148 skills in this repo. Your data stays on your computer, and you can optionally scale to cloud compute via [Modal](https://modal.com/) for heavy workloads. [Get started here.](https://github.com/K-Dense-AI/k-dense-byok)
+> **New: [K-Dense BYOK](https://github.com/K-Dense-AI/k-dense-byok)** — A free, open-source AI co-scientist that runs on your desktop, powered by Scientific Agent Skills. Bring your own API keys, pick from 40+ models, and get a full research workspace with web search, file handling, 100+ scientific databases, and access to all 150 skills in this repo. Your data stays on your computer, and you can optionally scale to cloud compute via [Modal](https://modal.com/) for heavy workloads. [Get started here.](https://github.com/K-Dense-AI/k-dense-byok)
 
 > **Stay up to date:** Follow K-Dense on [X](https://x.com/k_dense_ai), [LinkedIn](https://www.linkedin.com/company/k-dense-inc), and [YouTube](https://www.youtube.com/@K-Dense-Inc) for new skills, release announcements, walkthroughs, research workflow demos, and examples you can use with your own AI agent.
 
-A comprehensive collection of **148 ready-to-use scientific and research skills** (covering cancer genomics, drug-target binding, molecular dynamics, RNA velocity, geospatial science, time series forecasting, scientific ML resource discovery via Hugging Science, 78+ scientific databases, and more) for any AI agent that supports the open [Agent Skills](https://agentskills.io/) standard, created by [K-Dense](https://k-dense.ai). Works with **Cursor, Claude Code, Codex, Google Antigravity, and more**. Transform your AI agent into a research assistant capable of executing complex multi-step scientific workflows across biology, chemistry, medicine, and beyond.
+A comprehensive collection of **150 ready-to-use scientific and research skills** (covering cancer genomics, individual-level 1000 Genomes queries, hosted regulatory-sequence prediction, drug-target binding, molecular dynamics, RNA velocity, geospatial science, time series forecasting, scientific ML resource discovery via Hugging Science, 78+ scientific databases, and more) for any AI agent that supports the open [Agent Skills](https://agentskills.io/) standard, created by [K-Dense](https://k-dense.ai). Works with **Cursor, Claude Code, Codex, Google Antigravity, and more**. Transform your AI agent into a research assistant capable of executing complex multi-step scientific workflows across biology, chemistry, medicine, and beyond.
 
 > ⭐ **Help make AI for science easier to discover:** If Scientific Agent Skills saves you time, teaches your agent a workflow, or helps your lab move faster, please [star this repository](https://github.com/K-Dense-AI/scientific-agent-skills). A star is a public signal that these open, reusable research skills are worth maintaining: it helps scientists, engineers, and open-source contributors find the project, shows which agent-skill standards are gaining real adoption, and gives us a clear reason to keep expanding the collection for the community.
 
@@ -37,9 +37,9 @@ These skills enable your AI agent to seamlessly work with specialized scientific
 - 🧬 Bioinformatics & Genomics - Sequence analysis, single-cell RNA-seq, gene regulatory networks, variant annotation, phylogenetic analysis
 - 🧪 Cheminformatics & Drug Discovery - Molecular property prediction, virtual screening, ADMET analysis, molecular docking, lead optimization
 - 🔬 Proteomics & Mass Spectrometry - LC-MS/MS processing, peptide identification, spectral matching, protein quantification
-- 🏥 Clinical Research & Precision Medicine - Clinical trials, pharmacogenomics, variant interpretation, drug safety, clinical decision support, treatment planning
-- 🧠 Healthcare AI & Clinical ML - EHR analysis, physiological signal processing, medical imaging, clinical prediction models
-- 🖼️ Medical Imaging & Digital Pathology - DICOM processing, whole slide image analysis, computational pathology, radiology workflows
+- 🏥 Clinical Research & Evidence Workflows - Clinical trials, pharmacogenomics, variant evidence review, aggregate decision-support evaluation, source-bound draft report structures, and formatting of clinician-authored treatment decisions
+- 🧠 Healthcare AI & Biosignal Research - EHR and model research, physiological signal analysis, and retrospective validation—not patient-specific diagnosis, treatment, alarms, or deployment decisions
+- 🖼️ Medical Imaging & Digital Pathology - Privacy-aware DICOM processing and research-only whole-slide image analysis, computational pathology, and radiology data workflows
 - 🤖 Machine Learning & AI - Deep learning, reinforcement learning, time series analysis, model interpretability, Bayesian methods
 - 🔮 Materials Science & Chemistry - Crystal structure analysis, phase diagrams, metabolic modeling, computational chemistry
 - 🌌 Physics & Astronomy - Astronomical data analysis, coordinate transformations, cosmological calculations, symbolic mathematics, physics computations
@@ -47,11 +47,11 @@ These skills enable your AI agent to seamlessly work with specialized scientific
 - 📊 Data Analysis & Visualization - Statistical analysis, network analysis, time series, publication-quality figures, large-scale data processing, EDA
 - 🌍 Geospatial Science & Remote Sensing - Satellite imagery processing, GIS analysis, spatial statistics, terrain analysis, machine learning for Earth observation
 - 🧪 Laboratory Automation - Liquid handling protocols, lab equipment control, workflow automation, LIMS integration
-- 📚 Scientific Communication - Literature review, peer review, scientific writing, document processing, posters, slides, schematics, citation management
+- 📚 Scientific Communication - Evidence-traceable writing, confidential authorized peer review, literature synthesis, document processing, macro-free PPTX posters, slides, schematics, and citation management
 - 🔬 Multi-omics & Systems Biology - Multi-modal data integration, pathway analysis, network biology, systems-level insights
 - 🧬 Protein Engineering & Design - Protein language models, structure prediction, sequence design, function annotation
 - 🧰 Agent Platforms & Infrastructure - Build on Pi with SDK, RPC, extensions, custom providers/models, packages, TUI components, and session tooling
-- 🎓 Research Methodology - Hypothesis generation, scientific brainstorming, critical thinking, grant writing, scholar evaluation
+- 🎓 Research Methodology - Evidence-bounded candidate hypotheses, scientific brainstorming, critical thinking, grant writing, and qualitative low-stakes evaluation of scholarly works
 
 **Transform your AI coding agent into an 'AI Scientist' on your desktop!**
 
@@ -61,13 +61,13 @@ These skills enable your AI agent to seamlessly work with specialized scientific
 
 ## 📦 What's Included
 
-This repository provides **148 scientific and research skills** organized into the following categories:
+This repository provides **150 scientific and research skills** organized into the following categories:
 
 - **100+ Scientific & Financial Databases** - A unified database-lookup skill provides deterministic, provenance-rich access to 78 public databases (PubChem, ChEMBL, UniProt, COSMIC, ClinicalTrials.gov, FRED, USPTO, and more), plus dedicated skills for DepMap, Imaging Data Commons, PrimeKG, U.S. Treasury Fiscal Data, and Hugging Science (curated catalog of scientific datasets, models, and demos across 17 scientific domains on Hugging Face). Multi-database packages like BioServices (~40 bioinformatics services), BioPython (38 NCBI sub-databases via Entrez), and gget (20+ genomics databases) add further coverage
-- **70+ Optimized Python Package Skills** - Explicitly defined skills for RDKit, Scanpy, PyTorch Lightning, scikit-learn, BioPython, pyzotero, BioServices, PennyLane, Qiskit, Molecular Dynamics (OpenMM/MDAnalysis), scVelo, TimesFM, and others — with curated documentation, examples, and best practices. Note: the agent can write code using *any* Python package, not just these; these skills simply provide stronger, more reliable performance for the packages listed
+- **70+ Optimized Python Package Skills** - Explicitly defined, version-aware workflows for RDKit, Scanpy, PyTorch Lightning, scikit-learn, PyTDC, PathML, pydicom, NeuroKit2, PufferLib, QuTiP, GeoPandas, pymatgen, BioPython, Qiskit, Molecular Dynamics (OpenMM/MDAnalysis), and others. The agent can still use *any* Python package; these skills provide stronger, safer guidance for the packages listed
 - **9 Scientific Integration Skills** - Explicitly defined skills for Benchling, DNAnexus, LatchBio, OMERO, Protocols.io, Open Notebook, Ginkgo Cloud Lab, LabArchives, and Opentrons. Again, the agent is not limited to these — any API or platform reachable from Python is fair game; these skills are the optimized, pre-documented paths
-- **30+ Analysis & Communication Tools** - Literature review, scientific writing, peer review, document processing, Paperzilla, Exa Search, posters, slides, schematics, infographics, Mermaid diagrams, and more
-- **10+ Research & Clinical Tools** - Hypothesis generation, grant writing, clinical decision support, treatment plans, BIDS, regulatory compliance, scenario analysis, and workflow-derived skill drafting with Autoskill
+- **30+ Analysis & Communication Tools** - Literature review, evidence-traceable scientific writing, confidential peer review, document processing, Paperzilla, Exa Search, macro-free PPTX posters, slides, schematics, infographics, Mermaid diagrams, and more
+- **10+ Research & Clinical Tools** - Evidence-bounded hypothesis generation, grant writing, aggregate clinical decision-support research, clinician-authored treatment-plan formatting, BIDS, ISO 13485 evidence preparation, scenario analysis, and workflow-derived skill drafting with Autoskill
 
 Each skill includes:
 - ✅ Comprehensive documentation (`SKILL.md`)
@@ -106,13 +106,13 @@ Each skill includes:
 - **Multi-Step Workflows** - Execute complex pipelines with a single prompt
 
 ### 🎯 **Comprehensive Coverage**
-- **148 Skills** - Extensive coverage across all major scientific domains
+- **150 Skills** - Extensive coverage across all major scientific domains
 - **100+ Databases** - Unified access to 78+ databases via database-lookup, plus dedicated data access skills and multi-database packages like BioServices, BioPython, and gget
-- **70+ Optimized Python Package Skills** - RDKit, Scanpy, PyTorch Lightning, scikit-learn, BioServices, PennyLane, Qiskit, Molecular Dynamics (OpenMM/MDAnalysis), scVelo, TimesFM, and others (the agent can use any Python package; these are the pre-documented, higher-performing paths)
+- **70+ Optimized Python Package Skills** - Current, version-scoped guidance for packages including RDKit, Scanpy, PyTorch Lightning, scikit-learn, PyTDC, pydicom, PufferLib, QuTiP, GeoPandas, pymatgen, Qiskit, Molecular Dynamics (OpenMM/MDAnalysis), scVelo, and TimesFM (the agent can use any Python package; these are the pre-documented paths)
 
 ### 🔧 **Easy Integration**
 - **Simple Setup** - Copy skills to your skills directory and start working
-- **Automatic Discovery** - Your agent automatically finds and uses relevant skills
+- **Configured Discovery** - Compatible hosts can find and use relevant skills from their configured skill paths
 - **Well Documented** - Each skill includes examples, use cases, and best practices
 
 ### 🌟 **Maintained & Supported**
@@ -124,7 +124,7 @@ Each skill includes:
 
 ## 🎯 Getting Started
 
-### Option 1: npx (all platforms)
+### Option 1: npx (supported hosts)
 
 Install Scientific Agent Skills with a single command:
 
@@ -132,7 +132,7 @@ Install Scientific Agent Skills with a single command:
 npx skills add K-Dense-AI/scientific-agent-skills
 ```
 
-This is the official standard approach for installing Agent Skills across **all platforms**, including **Claude Code**, **Claude Cowork**, **Codex**, **Gemini CLI**, **Google Antigravity**, **Cursor**, **OpenClaw**, **NVIDIA NemoClaw**, **Hermes**, **Pi**, and any other agent that supports the open [Agent Skills](https://agentskills.io/) standard.
+This is a common standards-based installer for supported Agent Skills hosts, including current versions of **Claude Code**, **Claude Cowork**, **Codex**, **Gemini CLI**, **Google Antigravity**, and **Cursor**. Confirm installation paths and optional metadata behavior in your host's current documentation.
 
 ### Option 2: GitHub CLI (`gh skill`)
 
@@ -178,24 +178,24 @@ gh skill update --all
 
 ### Other Agent Skills hosts (OpenClaw, NemoClaw, Pi, Hermes, …)
 
-You usually don't need anything host-specific. `npx skills add` (Option 1) installs into the shared `~/.agents/skills/` convention, and any compliant client that scans that directory — including **OpenClaw**, **NVIDIA NemoClaw** (an OpenClaw-based secure runtime), and **Pi** — discovers the skills automatically. Project-scoped installs land in `.agents/skills/` and work the same way. To install without the CLI, clone straight into either location:
+Agent hosts differ in install paths, discovery settings, and support for optional frontmatter fields. `npx skills add` (Option 1) commonly installs into the `~/.agents/skills/` convention, with project-scoped installs under `.agents/skills/`; confirm both paths against your host's current documentation. To install manually on a host configured to scan one of those locations:
 
 ```bash
 git clone https://github.com/K-Dense-AI/scientific-agent-skills.git ~/.agents/skills/scientific-agent-skills   # user-level
 git clone https://github.com/K-Dense-AI/scientific-agent-skills.git .agents/skills/scientific-agent-skills      # project-level
 ```
 
-**Hermes** is the one host that uses its own registry instead of the shared directory, so add the repo as a tap:
+For Hermes versions that support skill taps, add the repository as a tap:
 
 ```bash
 hermes skills tap add K-Dense-AI/scientific-agent-skills
 ```
 
-These skills stay portable across all of them: `metadata` is single-line JSON (so OpenClaw's line-based reader parses it), credentialed skills declare a top-level `required_environment_variables` field (so Hermes prompts for keys), and unknown fields are ignored everywhere else. Because 148 skills add up to a lot of standing context, consider installing a topical subset rather than the whole collection.
+Every `SKILL.md` has YAML frontmatter, but legacy and community skills vary in `metadata` formatting (block or flow style) and optional extension fields. Repository updates must keep `metadata.version` as a quoted numeric string and pass canonical `skills-ref validate ./skills/<skill-name>` checks. Hosts may interpret optional metadata and credential prompts differently, so verify behavior on the target host. Because 150 skills add up to a lot of standing context, consider installing a topical subset rather than the whole collection.
 
 > **NemoClaw note:** NemoClaw runs agents inside NVIDIA OpenShell with default-deny outbound networking. Skills are discovered and loaded normally, but any skill that needs the network — package installs via `uv`, or API calls (Exa, Parallel, Benchling, NCBI, Materials Project, …) — only works once the operator pre-approves the relevant domains in the OpenShell TUI.
 
-**That's it!** Your AI agent will automatically discover the skills and use them when relevant to your scientific tasks. You can also invoke any skill manually by mentioning the skill name in your prompt.
+**That's it!** A compatible host can discover the skills from its configured paths and use them when relevant. You can also invoke any skill manually by mentioning the skill name in your prompt.
 
 ---
 
@@ -281,7 +281,7 @@ For more installation options and details, visit the [official uv documentation]
 Once you've installed the skills, you can ask your AI agent to execute complex multi-step scientific workflows. Here are some example prompts:
 
 ### 🧪 Drug Discovery Pipeline
-**Goal**: Find novel EGFR inhibitors for lung cancer treatment
+**Goal**: Prioritize EGFR inhibitor candidates for preclinical lung-cancer research
 
 **Prompt**:
 ```
@@ -340,18 +340,20 @@ MedChem/molfeat.
 
 ---
 
-### 🏥 Clinical Variant Interpretation
-**Goal**: Analyze VCF file for hereditary cancer risk assessment
+### 🏥 Research Variant Evidence Review
+**Goal**: Annotate a synthetic or properly de-identified VCF for hereditary-cancer research and qualified review
 
 **Prompt**:
 ```
-Use available skills you have access to whenever possible. Parse VCF with pysam, annotate variants with Ensembl VEP, query ClinVar for pathogenicity, 
-check COSMIC for cancer mutations, retrieve gene info from NCBI Gene, analyze protein impact 
-with UniProt, search PubMed for case reports, check ClinPGx for pharmacogenomics, generate 
-clinical report with document processing tools, and find matching trials on ClinicalTrials.gov.
+Use available skills you have access to whenever possible. Work only with authorized synthetic
+or de-identified data. Parse the VCF with pysam, annotate variants with Ensembl VEP, retrieve
+ClinVar/COSMIC/NCBI Gene/UniProt evidence, and verify literature sources. Build an evidence-
+traceable research summary with scientific-writing. If clinical-reports is used, create only a
+visibly marked draft structure from a verified source-fact manifest for qualified review; do not
+diagnose, assess individual risk, recommend treatment, or determine trial eligibility.
 ```
 
-**Skills Used**: pysam, database-lookup, paper-lookup, clinical-reports, docx, pdf
+**Skills Used**: pysam, database-lookup, paper-lookup, scientific-writing, clinical-reports
 
 ---
 
@@ -384,16 +386,19 @@ networks, and search GEO for similar patterns.
 ### 🧬 Bioinformatics & Genomics
 - **Sequence Analysis**: Process DNA/RNA/protein sequences with BioPython and pysam
 - **Single-Cell Analysis**: Analyze 10X Genomics data with Scanpy, identify cell types, infer GRNs with Arboreto
-- **Variant Annotation**: Annotate VCF files with Ensembl VEP, query ClinVar for pathogenicity
+- **Variant Annotation**: Annotate research VCF files with Ensembl VEP and retrieve ClinVar evidence for qualified interpretation
 - **Variant Database Management**: Build scalable VCF databases with TileDB-VCF for incremental sample addition, efficient population-scale queries, and compressed storage of genomic variant data
+- **Population Genomics**: Query variants, cohort sample IDs, and relatedness in the 3,202-person GRCh38 1000 Genomes cohort with OneKGPd
+- **Regulatory Sequence Models**: Run hosted Genomic Intelligence promoter, splice, enhancer, chromatin, expression, and gene-annotation predictions for research—not clinical or diagnostic decisions
 - **Gene Discovery**: Query NCBI Gene, UniProt, and Ensembl for comprehensive gene information
 - **Network Analysis**: Identify protein-protein interactions via STRING, map to pathways (KEGG, Reactome)
 
-### 🏥 Clinical Research & Precision Medicine
-- **Clinical Trials**: Search ClinicalTrials.gov for relevant studies, analyze eligibility criteria
-- **Variant Interpretation**: Annotate variants with ClinVar, COSMIC, and ClinPGx for pharmacogenomics
-- **Drug Safety**: Query FDA databases for adverse events, drug interactions, and recalls
-- **Precision Therapeutics**: Match patient variants to targeted therapies and clinical trials
+### 🏥 Clinical Research & Evidence Workflows
+- **Clinical Trials**: Analyze aggregate trial landscapes and protocol criteria without deciding individual eligibility
+- **Variant Evidence Review**: Annotate authorized research data with ClinVar, COSMIC, and ClinPGx; qualified professionals retain interpretation responsibility
+- **Drug Safety Research**: Query FDA databases for aggregate adverse-event, interaction, and recall evidence
+- **Decision-Support Evaluation**: Prepare synthetic or aggregate evaluation, evidence-profile, privacy, and governance artifacts—not live clinical decisions
+- **Clinician-Authored Documentation**: Structure verified source-bound report drafts and format treatment decisions already made by authorized licensed professionals
 
 ### 🔬 Multi-Omics & Systems Biology
 - **Multi-Omics Integration**: Combine RNA-seq, proteomics, and metabolomics data
@@ -405,28 +410,29 @@ networks, and search GEO for similar patterns.
 - **Statistical Analysis**: Perform hypothesis testing, power analysis, and experimental design
 - **Publication Figures**: Create publication-quality visualizations with matplotlib and seaborn
 - **Network Visualization**: Visualize biological networks with NetworkX
-- **Report Generation**: Generate comprehensive reports with the PDF, DOCX, PPTX, XLSX, MarkItDown, LiteParse, and clinical-reporting skills
+- **Report Generation**: Produce evidence-traceable research reports with Scientific Writing and document tools; clinical-reporting outputs remain visibly marked drafts built only from verified synthetic, de-identified, or aggregate source facts
 
 ### 🧪 Laboratory Automation
-- **Protocol Design**: Create Opentrons protocols for automated liquid handling
-- **LIMS Integration**: Integrate with Benchling and LabArchives for data management
-- **Workflow Automation**: Automate multi-step laboratory workflows
+- **Protocol Design**: Author and simulate Opentrons or PyLabRobot protocols before trained-operator review
+- **LIMS/ELN Integration**: Prepare scoped Benchling and LabArchives operations with explicit authorization for remote writes
+- **Workflow Automation**: Validate and simulate multi-step laboratory workflows offline; physical execution stays behind equipment-specific operator safety gates
 
 ---
 
 ## 📚 Available Skills
 
-This repository contains **148 scientific and research skills** organized across multiple domains. Each skill provides comprehensive documentation, code examples, and best practices for working with scientific libraries, databases, and tools.
+This repository contains **150 scientific and research skills** organized across multiple domains. Each skill provides comprehensive documentation, code examples, and best practices for working with scientific libraries, databases, and tools.
 
 ### Skill Categories
 
 > **Note:** The Python package and integration skills listed below are *explicitly defined* skills — curated with documentation, examples, and best practices for stronger, more reliable performance. They are not a ceiling: the agent can install and use *any* Python package or call *any* API, even without a dedicated skill. The skills listed simply make common workflows faster and more dependable.
 
-#### 🧬 **Bioinformatics & Genomics** (23 skills)
+#### 🧬 **Bioinformatics & Genomics** (25 skills)
 - RNA-seq pipelines: Bulk RNA-seq (end-to-end FASTQ -> counts -> DE -> enrichment orchestrator)
 - Sequence analysis: BioPython, pysam, scikit-bio, BioServices
 - Single-cell analysis: Scanpy, AnnData, scvi-tools, scVelo (RNA velocity), Arboreto, Cellxgene Census
-- Genomic tools: gget, geniml, gtars, deepTools, FlowIO, Polars-Bio, Zarr, TileDB-VCF
+- Genomic tools: gget, current geniml/Gtars interval workflows, deepTools, FlowIO, Polars-Bio, Zarr, TileDB-VCF
+- Population and sequence intelligence: OneKGPd (individual-level 1000 Genomes cohort queries) and Genomic Intelligence (hosted regulatory/gene-expression predictions; research only)
 - Differential expression: PyDESeq2
 - Functional enrichment: Pathway Enrichment (ORA, GSEA/preranked, ssGSEA via gseapy + g:Profiler; GO, KEGG, Reactome, WikiPathways, MSigDB)
 - Phylogenetics: ETE Toolkit, Phylogenetics (MAFFT, IQ-TREE 2, FastTree)
@@ -438,29 +444,31 @@ This repository contains **148 scientific and research skills** organized across
 - Molecular dynamics: OpenMM + MDAnalysis (MD simulation & trajectory analysis)
 - Cloud quantum chemistry: Rowan (pKa, docking, cofolding)
 - Drug-likeness: MedChem
-- Benchmarks: PyTDC
+- Benchmarks: PyTDC 1.1.15 on its verified CPython 3.11 compatibility stack
 
 #### 🔬 **Proteomics & Mass Spectrometry** (2 skills)
 - Spectral processing: matchms, pyOpenMS
 
-#### 🏥 **Clinical Research & Precision Medicine** (8 skills)
+#### 🏥 **Clinical Research & Evidence Workflows** (7 skills)
 - Clinical databases: via Database Lookup (ClinicalTrials.gov, ClinVar, ClinPGx, COSMIC, FDA, cBioPortal, Monarch, and more)
 - Cancer genomics: DepMap (cancer dependency scores, drug sensitivity)
 - Cancer imaging: Imaging Data Commons (NCI radiology & pathology datasets via idc-index)
-- Healthcare AI: PyHealth, NeuroKit2, Clinical Decision Support
-- Clinical documentation: Clinical Reports, Treatment Plans
+- Healthcare AI research: PyHealth
+- Decision-support research: local, aggregate or synthetic Clinical Decision Support evaluation and governance artifacts only
+- Clinical documentation: source-bound Clinical Reports drafts and formatting of verified clinician-authored decisions with Treatment Plans; neither skill diagnoses or recommends care
 
 #### 🖼️ **Medical Imaging & Digital Pathology** (3 skills)
-- DICOM processing: pydicom
-- Whole slide imaging: histolab, PathML
+- DICOM processing: pydicom 3.0.2 with privacy-first local preflight and no diagnostic or de-identification-compliance claims
+- Whole slide imaging: histolab and research-only PathML 3.0.5
 
-#### 🧠 **Neuroscience & Electrophysiology** (2 skills)
+#### 🧠 **Neuroscience & Electrophysiology** (3 skills)
 - Data standards: BIDS (Brain Imaging Data Structure for neuroscience and biomedical datasets)
 - Neural recordings: Neuropixels-Analysis (extracellular spikes, silicon probes, spike sorting)
+- Physiological signals: NeuroKit2 0.2.13 for reproducible research workflows—not diagnosis, monitoring decisions, or medical-device validation
 
 #### 🤖 **Machine Learning & AI** (14 core skills)
-- Deep learning: PyTorch Lightning, Transformers, Stable Baselines3, PufferLib
-- Classical ML: scikit-learn, scikit-survival, SHAP
+- Deep learning: PyTorch Lightning, Transformers, Stable Baselines3, and version-separated PufferLib 3.0/4.0 workflows
+- Classical ML: scikit-learn, scikit-survival 0.28, and SHAP
 - Time series: aeon, TimesFM (Google's zero-shot foundation model for univariate forecasting)
 - Bayesian methods: PyMC
 - Optimization: PyMOO
@@ -469,39 +477,38 @@ This repository contains **148 scientific and research skills** organized across
 - Statistical modeling: statsmodels
 
 #### 🔮 **Materials Science, Chemistry & Physics** (7 skills)
-- Materials: Pymatgen
+- Materials: current split pymatgen wrapper/core plus explicitly bounded Materials Project queries
 - Metabolic modeling: COBRApy
 - Astronomy: Astropy
-- Quantum computing: Cirq, PennyLane, Qiskit, QuTiP
+- Quantum computing: Cirq, PennyLane, Qiskit, QuTiP 5.3
 
 #### ⚙️ **Engineering & Simulation** (4 skills)
-- Numerical computing: MATLAB/Octave
-- Computational fluid dynamics: FluidSim
-- Discrete-event simulation: SimPy
+- Numerical computing: proprietary MATLAB R2026a and distinct GNU Octave 11.3 planning/review workflows
+- Computational fluid dynamics: bounded FluidSim 0.9 simulations with numerical-validity and HPC checks
+- Discrete-event simulation: SimPy 4.1.2 with replication, warm-up, and output-analysis guidance
 - Symbolic math: SymPy
 
 #### 📊 **Data Analysis & Visualization** (21 skills)
 - Visualization: Matplotlib, Seaborn, Scientific Visualization
-- Geospatial analysis: GeoPandas, GeoMaster (remote sensing, GIS, satellite imagery, spatial ML, 500+ examples)
+- Geospatial analysis: GeoPandas 1.1.4 and GeoMaster (remote sensing, GIS, satellite imagery, spatial ML, 500+ examples)
 - Data processing: Dask, Polars, Vaex
 - Network analysis: NetworkX
 - Document processing: LiteParse (local PDF/document parsing with bounding boxes and OCR), MarkItDown, PDF, DOCX, PPTX, and XLSX
 - Infographics: Infographics (AI-powered professional infographic creation)
 - Diagrams: Markdown & Mermaid Writing (text-based diagrams as default documentation standard)
-- Exploratory data analysis: EDA workflows
+- Exploratory data analysis: bounded local EDA for explicitly supported formats, with unknown formats failing closed
 - Statistical analysis: Statistical Analysis workflows
 - Experimental design: Experimental Design (randomization, blocking, factorial/fractional-factorial DOE, crossover, cluster, sequential designs; pyDOE3)
 - Statistical power: Statistical Power (sample-size & power for t-tests, ANOVA, proportions, correlation, regression — closed-form plus simulation-based for GLMs, mixed models, and cluster designs)
 
 #### 🧪 **Laboratory Automation** (6 skills)
-- Liquid handling: PyLabRobot and Opentrons
+- Liquid handling: offline-first PyLabRobot planning/simulation and Opentrons authoring, with physical execution behind explicit operator safety gates
 - Cloud lab: Ginkgo Cloud Lab (protein expression & purification across cell-free/E. coli/Pichia, IVT RNA synthesis, thermal shift and Echo-MS assays, SPR onboarding, fluorescent pixel art via autonomous RAC infrastructure)
-- Protocol management: Protocols.io
-- LIMS integration: Benchling, LabArchives
+- Protocol management: bounded protocols.io reads across documented v3/v4 endpoints and non-executing write plans
+- LIMS/ELN integration: Benchling and the separate LabArchives legacy ELN and Inventory v1 APIs
 
-#### 🔬 **Multi-omics & Systems Biology** (4 skills)
+#### 🔬 **Multi-omics & Systems Biology** (3 skills)
 - Pathway analysis: via Database Lookup (KEGG, Reactome, STRING) and PrimeKG
-- Multi-omics: HypoGeniC
 - Data management: LaminDB
 
 #### 🧬 **Protein Engineering & Design** (4 skills)
@@ -515,10 +522,10 @@ This repository contains **148 scientific and research skills** organized across
 - Advanced paper search: BGPT Paper Search (25+ structured fields per paper — methods, results, sample sizes, quality scores — from full text, not just abstracts)
 - Web intelligence: Parallel Web (web search, URL/PDF extraction, deep research, structured enrichment, entity discovery, and recurring monitoring), Exa Search, and Research Lookup
 - Research notebooks: Open Notebook (self-hosted NotebookLM alternative — PDFs, videos, audio, web pages; 16+ AI providers; multi-speaker podcast generation)
-- Writing: Scientific Writing, Peer Review
+- Writing: evidence-traceable Scientific Writing and local, confidential, authorized Peer Review
 - Document processing: LiteParse, PDF, DOCX, PPTX, XLSX, and MarkItDown
 - Publishing and paper workflows: Venue Templates
-- Presentations: Scientific Slides, LaTeX Posters, PPTX Posters
+- Presentations: Scientific Slides, LaTeX Posters, and macro-free PPTX Posters generated from author-approved local manifests
 - Diagrams: Scientific Schematics, Markdown & Mermaid Writing
 - Infographics: Infographics (10 types, 8 styles, colorblind-safe palettes)
 - Citations: Citation Management, pyzotero
@@ -540,23 +547,24 @@ This repository contains **148 scientific and research skills** organized across
 - Workflow engines: Nextflow (build/run/debug Nextflow & nf-core pipelines — DSL2 modules, executors/containers, HPC/cloud scaling) and pacsomatic (operator toolkit for the nf-core/pacsomatic tumor-normal somatic variant-calling workflow)
 - Microscopy: OMERO
 - Automation: Opentrons
-- Resource detection: Get Available Resources
+- Resource detection: Get Available Resources on request or before a clearly resource-sensitive local workload; redacted and without stress tests
 - Workflow mining: Autoskill (local screenpipe-based repeated workflow detection and skill drafting)
 - Agent platform development: Pi Agent (using Pi as a terminal coding harness and building on it with SDK, RPC/JSONL, extensions, custom providers/models, packages, TUI components, and session tooling)
 
-#### 🎓 **Research Methodology & Planning** (12 skills)
-- Ideation: Scientific Brainstorming, Hypothesis Generation
+#### 🎓 **Research Methodology & Planning** (13 skills)
+- Ideation: evidence-aware Scientific Brainstorming and non-scoring Hypothesis Generation that keeps hypotheses labeled as candidates
+- Text-dataset hypothesis software: HypoGeniC/HypoRefine produces candidate textual patterns and task-prediction statistics, not validated scientific hypotheses
 - Autonomous optimization: Arbor (Hypothesis Tree Refinement — iteratively improve a code/model/agent-harness/data artifact against a dev evaluator while a held-out test gate guards against overfitting)
-- Critical analysis: Scientific Critical Thinking, Scholar Evaluation
+- Critical analysis: Scientific Critical Thinking and qualitative, low-stakes Scholar Evaluation of works—never ranking people or supporting consequential decisions
 - Scenario analysis: What-If Oracle (4–6 branch possibility exploration, contingency planning, decision stress-testing)
 - Multi-perspective deliberation: Consciousness Council (diverse expert viewpoints, devil's advocate analysis)
 - Cognitive profiling: DHDNA Profiler (extract thinking patterns and cognitive signatures from any text)
 - Funding: Research Grants
 - Discovery: Research Lookup, Paper Lookup (10 academic databases)
-- Market analysis: Market Research Reports
+- Market analysis: evidence-traceable Market Research Reports with assumption-led sizing and forecast sensitivity
 
 #### ⚖️ **Regulatory & Standards** (1 skill)
-- Medical device standards: ISO 13485 Certification
+- Medical device standards: draft ISO 13485 QMS evidence-preparation artifacts for qualified review—not compliance, audit, or certification decisions
 
 > 📖 **For complete details on all skills**, see [docs/skills.md](docs/skills.md)
 
@@ -690,7 +698,7 @@ A: No. Each skill has its own license specified in the `license` metadata field
 A: We regularly update skills to reflect the latest versions of packages and APIs. Major updates are announced in release notes.
 
 **Q: Can I use this with other AI models?**  
-A: The skills follow the open [Agent Skills](https://agentskills.io/) standard and work with any compatible agent, including Cursor, Claude Code, Codex, Google Antigravity, OpenClaw, NVIDIA NemoClaw, Hermes, and Pi.
+A: The core `SKILL.md` format follows the open [Agent Skills](https://agentskills.io/) standard. Installation paths, discovery, and optional metadata support vary by host and version, so confirm your target host's current documentation.
 
 ### Installation & Setup
 
@@ -745,7 +753,7 @@ Recommended practice:
   title = {Scientific Agent Skills: A Comprehensive Collection of Scientific Tools for AI Agents},
   year = {2026},
   url = {https://github.com/K-Dense-AI/scientific-agent-skills},
-  note = {148 skills covering databases, packages, integrations, and analysis tools}
+  note = {150 skills covering databases, packages, integrations, and analysis tools}
 }
 ```
 

docs/open-source-sponsors.md

@@ -1,6 +1,6 @@
 # Support the Open Source Projects We Depend On
 
-Scientific Agent Skills is built on the shoulders of giants. The 148 skills in this repository leverage dozens of incredible open source projects created and maintained by dedicated developers and research communities around the world.
+Scientific Agent Skills is built on the shoulders of giants. The 150 skills in this repository leverage dozens of incredible open source projects created and maintained by dedicated developers and research communities around the world.
 
 **If you find value in these skills, please consider supporting the underlying open source projects that make them possible.**
 
@@ -31,6 +31,8 @@ Scientific Agent Skills is built on the shoulders of giants. The 148 skills in t
 | **scikit-bio** | Bioinformatics library | [GitHub](https://github.com/scikit-bio/scikit-bio) |
 | **gget** | Gene and transcript info retrieval | [GitHub](https://github.com/pachterlab/gget) |
 | **deepTools** | Tools for deep-sequencing data | [GitHub](https://github.com/deeptools/deepTools) |
+| **geniml** | Machine learning for genomic interval sets | [GitHub](https://github.com/databio/geniml) |
+| **Gtars** | Rust/Python genomic interval toolkit | [GitHub](https://github.com/databio/gtars) |
 | **ETE Toolkit** | Phylogenetic tree analysis | [GitHub](https://github.com/etetoolkit/ete) |
 
 ### Cheminformatics & Drug Discovery
@@ -63,10 +65,12 @@ Scientific Agent Skills is built on the shoulders of giants. The 148 skills in t
 | **PyMC** | Probabilistic programming | [GitHub](https://github.com/pymc-devs/pymc) - [Donate](https://numfocus.org/donate-to-pymc) |
 | **SHAP** | Model interpretability | [GitHub](https://github.com/shap/shap) |
 | **Stable Baselines3** | Reinforcement learning | [GitHub](https://github.com/DLR-RM/stable-baselines3) |
+| **PufferLib** | High-performance reinforcement learning environments and training | [GitHub](https://github.com/PufferAI/PufferLib) |
 | **scikit-survival** | Survival analysis | [GitHub](https://github.com/sebp/scikit-survival) |
 | **aeon** | Time series ML toolkit | [GitHub](https://github.com/aeon-toolkit/aeon) |
 | **PyMOO** | Multi-objective optimization | [GitHub](https://github.com/anyoptimization/pymoo) |
 | **UMAP** | Dimensionality reduction | [GitHub](https://github.com/lmcinnes/umap) |
+| **HypoGeniC** | LLM-assisted candidate textual hypothesis generation from labeled data | [GitHub](https://github.com/ChicagoHAI/hypothesis-generation) |
 
 ### Data Science & Visualization
 
@@ -90,6 +94,7 @@ Scientific Agent Skills is built on the shoulders of giants. The 148 skills in t
 | **pydicom** | DICOM file handling | [GitHub](https://github.com/pydicom/pydicom) |
 | **histolab** | Digital pathology preprocessing | [GitHub](https://github.com/histolab/histolab) |
 | **PathML** | Pathology ML toolkit | [GitHub](https://github.com/Dana-Farber-AIOS/pathml) |
+| **OMERO.py** | Python client for the OMERO microscopy data platform | [GitHub](https://github.com/ome/omero-py) |
 
 ### Healthcare & Clinical
 
@@ -119,8 +124,9 @@ Scientific Agent Skills is built on the shoulders of giants. The 148 skills in t
 
 | Project | Description | Links |
 |---------|-------------|-------|
-| **SimPy** | Discrete-event simulation | [GitHub](https://github.com/TeamSim/SimPy) |
+| **SimPy** | Discrete-event simulation | [GitLab](https://gitlab.com/team-simpy/simpy/) |
 | **FluidSim** | CFD framework | [GitHub](https://github.com/fluiddyn/fluidsim) |
+| **FluidFFT** | FFT framework and plugins used by FluidSim | [GitHub](https://github.com/fluiddyn/fluidfft) |
 
 ### Laboratory & Automation
 
@@ -141,6 +147,12 @@ Scientific Agent Skills is built on the shoulders of giants. The 148 skills in t
 | **Zarr** | Chunked array storage | [GitHub](https://github.com/zarr-developers/zarr-python) |
 | **FlowIO** | Flow cytometry I/O | [GitHub](https://github.com/whitews/FlowIO) |
 
+### Scientific Communication & Documents
+
+| Project | Description | Links |
+|---------|-------------|-------|
+| **python-pptx** | Create and inspect editable PowerPoint `.pptx` files | [GitHub](https://github.com/scanny/python-pptx) |
+
 ---
 
 ## NumFOCUS-Sponsored Projects

docs/skills.md

@@ -2,221 +2,223 @@
 
 ## Scientific Databases & Data Access
 
-- **Database Lookup** - Deterministically query 78 public scientific, biomedical, materials science, regulatory, finance, and demographics databases through documented REST APIs. Adds retrieval contracts, endpoint provenance, pagination/count reconciliation, rate-limit guidance, and untrusted-response handling for reproducible lookups of compounds, drugs, proteins, genes, pathways, variants, clinical trials, patents, SEC filings, economic indicators, structures, astronomical objects, earthquakes, weather, and related database-backed scientific facts
-- **DepMap** - Query the Cancer Dependency Map (DepMap) for cancer cell line gene dependency scores (CRISPR Chronos), drug sensitivity data, and gene effect profiles. Use for identifying cancer-specific vulnerabilities, synthetic lethal interactions, and validating oncology drug targets
-- **Imaging Data Commons** - Query and download public cancer imaging data from NCI Imaging Data Commons using idc-index. Use for accessing large-scale radiology (CT, MR, PET) and pathology datasets for AI training or research. No authentication required. Query by metadata, visualize in browser, check licenses
-- **PrimeKG** - Query the Precision Medicine Knowledge Graph (PrimeKG) for multiscale biological data including genes, drugs, diseases, phenotypes, and more. Integrates 20+ biomedical resources into a single knowledge graph for drug repurposing, disease mechanism exploration, and target identification
-- **U.S. Treasury Fiscal Data** - Free, open REST API from the U.S. Department of the Treasury providing 54 datasets and 179 data tables covering federal fiscal data. No API key required. Access national debt (Debt to the Penny back to 1993, Historical Debt back to 1790), Daily Treasury Statements (TGA balances, deposits/withdrawals), Monthly Treasury Statements (federal budget receipts and outlays), Treasury securities auctions data (bills, notes, bonds, TIPS, FRNs since 1979), average interest rates on Treasury securities, Treasury reporting exchange rates (quarterly for 170+ currencies), I Bond and savings bond rates, TIPS/CPI data, and more. Supports filtering, sorting, pagination, and CSV/XML/JSON output formats
-- **Hugging Science** - Curated, LLM-friendly catalog of scientific datasets, models, blog posts, and interactive Spaces hosted on Hugging Face, spanning 17 scientific domains (astronomy, benchmark, biology, biotechnology, chemistry, climate, conservation, earth-science, ecology, energy, engineering, genomics, materials-science, mathematics, medicine, physics, scientific-reasoning). Discovery happens via huggingscience.co (with `llms.txt`, `llms-full.txt`, and per-topic markdown files designed for agent consumption); usage goes through standard Hugging Face APIs. Includes a bundled `fetch_catalog.py` script for filtered access by topic, type (datasets/models/blogs/spaces), or free-text search, plus reference guides for loading datasets via the `datasets` library (with streaming for billion-token corpora), running models via `transformers` or the Inference API/Providers (handling `trust_remote_code` requirements for custom architectures like Evo-2), and calling Spaces via `gradio_client` (with a worked BoltzGen example for protein binder design). Authenticates gated resources via `HF_TOKEN` from `.env`. Use cases: discovering the right dataset/model for a scientific ML task without trawling the broader Hub, fine-tuning on curated scientific data, citing methodology blogs from dataset/model authors, running interactive scientific demos (binder design, theorem proving, weather modeling) without local GPU setup, and bridging from "I need a model for protein/genome/molecule/climate/materials/astronomy" to working code
+- **[Database Lookup](../skills/database-lookup/)** - Deterministically query 78 public scientific, biomedical, materials science, regulatory, finance, and demographics databases through documented REST APIs. Adds retrieval contracts, endpoint provenance, pagination/count reconciliation, rate-limit guidance, and untrusted-response handling for reproducible lookups of compounds, drugs, proteins, genes, pathways, variants, clinical trials, patents, SEC filings, economic indicators, structures, astronomical objects, earthquakes, weather, and related database-backed scientific facts
+- **[DepMap](../skills/depmap/)** - Query the Cancer Dependency Map (DepMap) for cancer cell line gene dependency scores (CRISPR Chronos), drug sensitivity data, and gene effect profiles. Use for identifying cancer-specific vulnerabilities, synthetic lethal interactions, and validating oncology drug targets
+- **[Imaging Data Commons](../skills/imaging-data-commons/)** - Query and download public cancer imaging data from NCI Imaging Data Commons using idc-index. Use for accessing large-scale radiology (CT, MR, PET) and pathology datasets for AI training or research. No authentication required. Query by metadata, visualize in browser, check licenses
+- **[PrimeKG](../skills/primekg/)** - Query the Precision Medicine Knowledge Graph (PrimeKG) for multiscale biological data including genes, drugs, diseases, phenotypes, and more. Integrates 20+ biomedical resources into a single knowledge graph for drug repurposing, disease mechanism exploration, and target identification
+- **[U.S. Treasury Fiscal Data](../skills/usfiscaldata/)** - Free, open REST API from the U.S. Department of the Treasury providing 54 datasets and 179 data tables covering federal fiscal data. No API key required. Access national debt (Debt to the Penny back to 1993, Historical Debt back to 1790), Daily Treasury Statements (TGA balances, deposits/withdrawals), Monthly Treasury Statements (federal budget receipts and outlays), Treasury securities auctions data (bills, notes, bonds, TIPS, FRNs since 1979), average interest rates on Treasury securities, Treasury reporting exchange rates (quarterly for 170+ currencies), I Bond and savings bond rates, TIPS/CPI data, and more. Supports filtering, sorting, pagination, and CSV/XML/JSON output formats
+- **[Hugging Science](../skills/hugging-science/)** - Curated, LLM-friendly catalog of scientific datasets, models, blog posts, and interactive Spaces hosted on Hugging Face, spanning 17 scientific domains (astronomy, benchmark, biology, biotechnology, chemistry, climate, conservation, earth-science, ecology, energy, engineering, genomics, materials-science, mathematics, medicine, physics, scientific-reasoning). Discovery happens via huggingscience.co (with `llms.txt`, `llms-full.txt`, and per-topic markdown files designed for agent consumption); usage goes through standard Hugging Face APIs. Includes a bundled `fetch_catalog.py` script for filtered access by topic, type (datasets/models/blogs/spaces), or free-text search, plus reference guides for loading datasets via the `datasets` library (with streaming for billion-token corpora), running models via `transformers` or the Inference API/Providers (handling `trust_remote_code` requirements for custom architectures like Evo-2), and calling Spaces via `gradio_client` (with a worked BoltzGen example for protein binder design). Authenticates gated resources via `HF_TOKEN` from `.env`. Use cases: discovering the right dataset/model for a scientific ML task without trawling the broader Hub, fine-tuning on curated scientific data, citing methodology blogs from dataset/model authors, running interactive scientific demos (binder design, theorem proving, weather modeling) without local GPU setup, and bridging from "I need a model for protein/genome/molecule/climate/materials/astronomy" to working code
 
 ## Scientific Integrations
 
 ### Laboratory Information Management Systems (LIMS) & R&D Platforms
-- **Benchling Integration** - Toolkit for integrating with Benchling's R&D platform, providing programmatic access to laboratory data management including registry entities (DNA sequences, proteins), inventory systems (samples, containers, locations), electronic lab notebooks (entries, protocols), workflows (tasks, automation), and data exports using Python SDK and REST API
+- **[Benchling Integration](../skills/benchling-integration/)** - Toolkit for integrating with Benchling's R&D platform, providing programmatic access to laboratory data management including registry entities (DNA sequences, proteins), inventory systems (samples, containers, locations), electronic lab notebooks (entries, protocols), workflows (tasks, automation), and data exports using Python SDK and REST API
 
 ### Cloud Platforms for Genomics & Biomedical Data
-- **DNAnexus Integration** - Build and operate reproducible genomics workloads on DNAnexus with current `dx`/`dxpy` APIs. Covers secure authentication and project context, resumable data transfers, safe metadata/archival/deletion operations, Ubuntu 24.04 app and applet development, modern region-scoped `dxapp.json` configuration, jobs and workflow analyses with cost/retry/resource controls, native workflows, WDL/CWL through dxCompiler, and licensed Nextflow imports. Includes offline validators for manifests and installed SDK compatibility
+- **[DNAnexus Integration](../skills/dnanexus-integration/)** - Build and operate reproducible genomics workloads on DNAnexus with current `dx`/`dxpy` APIs. Covers secure authentication and project context, resumable data transfers, safe metadata/archival/deletion operations, Ubuntu 24.04 app and applet development, modern region-scoped `dxapp.json` configuration, jobs and workflow analyses with cost/retry/resource controls, native workflows, WDL/CWL through dxCompiler, and licensed Nextflow imports. Includes offline validators for manifests and installed SDK compatibility
 
 ### Laboratory Automation
-- **Opentrons Integration** - Author, review, migrate, simulate, and troubleshoot current Opentrons Python Protocol API v2 workflows for Flex and OT-2. Covers robot-specific API versioning, deck fixtures, labware and adapters, pipettes and partial tip pickup, contamination-aware liquid handling, runtime parameters, liquid classes and sensing, Gripper moves, Thermocycler/Temperature/Heater-Shaker/Magnetic hardware, the Absorbance Plate Reader, Flex Stacker, App analysis, dry runs, and production safety checks. Includes pinned simulation dependencies and verified Flex 2.29 and OT-2 2.28 templates
-- **Ginkgo Cloud Lab** - Submit and manage protocols on Ginkgo Bioworks Cloud Lab (cloud.ginkgo.bio), a web-based interface for autonomous lab execution on Reconfigurable Automation Carts (RACs). Covers the full catalog: protein expression and purification across cell-free, E. coli, and Pichia systems (HiBiT/luminescence, A280 yield, and LabChip purity/size tiers, plus minibinder screening), cell-free validation ($39) and 24-condition DoE optimization ($199), IVT mRNA/circRNA synthesis with qPCR, thermal shift / developability assays, Echo-MS enzyme detection and analyte method onboarding, SPR target onboarding for kinetics, and fluorescent pixel art ($25/plate, 7-color E. coli palette). Includes the EstiMate AI agent for custom protocol feasibility and pricing
+- **[Opentrons Integration](../skills/opentrons-integration/)** - Author, review, migrate, simulate, and troubleshoot current Opentrons Python Protocol API v2 workflows for Flex and OT-2. Covers robot-specific API versioning, deck fixtures, labware and adapters, pipettes and partial tip pickup, contamination-aware liquid handling, runtime parameters, liquid classes and sensing, Gripper moves, Thermocycler/Temperature/Heater-Shaker/Magnetic hardware, the Absorbance Plate Reader, Flex Stacker, App analysis, dry runs, and production safety checks. Includes pinned simulation dependencies and verified Flex 2.29 and OT-2 2.28 templates
+- **[Ginkgo Cloud Lab](../skills/ginkgo-cloud-lab/)** - Submit and manage protocols on Ginkgo Bioworks Cloud Lab (cloud.ginkgo.bio), a web-based interface for autonomous lab execution on Reconfigurable Automation Carts (RACs). Covers the full catalog: protein expression and purification across cell-free, E. coli, and Pichia systems (HiBiT/luminescence, A280 yield, and LabChip purity/size tiers, plus minibinder screening), cell-free validation ($39) and 24-condition DoE optimization ($199), IVT mRNA/circRNA synthesis with qPCR, thermal shift / developability assays, Echo-MS enzyme detection and analyte method onboarding, SPR target onboarding for kinetics, and fluorescent pixel art ($25/plate, 7-color E. coli palette). Includes the EstiMate AI agent for custom protocol feasibility and pricing
 
 ### Electronic Lab Notebooks (ELN)
-- **LabArchives Integration** - Toolkit for interacting with LabArchives Electronic Lab Notebook (ELN) REST API. Provides programmatic access to notebooks (backup, retrieval, management), entries (creation, comments, attachments), user authentication, site reports and analytics, and third-party integrations (Protocols.io, GraphPad Prism, SnapGene, Geneious, Jupyter, REDCap). Includes Python scripts for configuration setup, notebook operations, and entry management. Supports multi-regional API endpoints (US, UK, Australia) and OAuth authentication
-- **Open Notebook** - Self-hosted, open-source alternative to Google NotebookLM for AI-powered research and document analysis. Organizes research materials into notebooks, ingests diverse content sources (PDFs, videos, audio, web pages, Office documents), generates AI-powered notes and summaries, creates multi-speaker podcasts from research, enables document chat with context-aware AI, and searches across materials with full-text and vector search. Supports 16+ AI providers including OpenAI, Anthropic, Google, Ollama, Groq, and Mistral with complete data privacy through self-hosting
+- **[LabArchives Integration](../skills/labarchive-integration/)** - Securely integrate with the separate official legacy ELN REST-like API and Inventory API v1. Covers signed requests, user authorization and UID flows, local LA-container validation, and the five documented ELN regions (US/rest of world, Australia/New Zealand, UK, Europe outside the UK, and Canada). Bundled Python 3.11+ tools read only named `LABARCHIVES_*` variables; remote writes require a reviewed, redacted plan and explicit approval
+- **[Open Notebook](../skills/open-notebook/)** - Self-hosted, open-source alternative to Google NotebookLM for AI-powered research and document analysis. Organizes research materials into notebooks, ingests diverse content sources (PDFs, videos, audio, web pages, Office documents), generates AI-powered notes and summaries, creates multi-speaker podcasts from research, enables document chat with context-aware AI, and searches across materials with full-text and vector search. Supports 16+ AI providers including OpenAI, Anthropic, Google, Ollama, Groq, and Mistral with complete data privacy through self-hosting
 
 ### Workflow Platforms & Cloud Execution
-- **LatchBio Integration** - Build, register, debug, and operate bioinformatics workflows on Latch with the current Python SDK and CLI. Covers typed workflow/task graphs, modern `LPath` data operations plus `LatchFile`/`LatchDir`, transactional Registry APIs and samplesheets, CPU/GPU resource sizing, UI metadata and launch plans, staging and programmatic execution, Nextflow and version-aware Snakemake packaging, ready-to-use workflow discovery, automations, and the OAuth-based Latch MCP
-- **Nextflow** - Build, run, and debug Nextflow data pipelines and nf-core workflows end to end. Covers writing and testing DSL2 modules/subworkflows (processes, channels, operators, nf-test), running community pipelines (nf-core/rnaseq, nf-core/sarek), authoring samplesheets and `nextflow.config`, configuring executors and containers (Docker, Singularity/Apptainer, Conda, Wave), scaling to HPC/SLURM or cloud (AWS Batch, Google Batch, Azure, Kubernetes), and debugging failed or `-resume` runs. Use for any reproducible scientific/bioinformatics workflow work and for authoring nf-core-compliant pipelines, modules, configs, and linting
-- **pacsomatic** - Operator toolkit for running the nf-core/pacsomatic matched tumor-normal (somatic variant calling) workflow from BAM inputs. Validates run inputs, generates pacsomatic-compliant samplesheets, prepares reproducible Nextflow launch artifacts, runs locally or submits to schedulers (LSF/Slurm/PBS/SGE), and triages execution and startup failures. Use to prepare launch commands/scripts, perform dry-run checks, or troubleshoot pipeline and scheduler submission errors
+- **[LatchBio Integration](../skills/latchbio-integration/)** - Build, register, debug, and operate bioinformatics workflows on Latch with the current Python SDK and CLI. Covers typed workflow/task graphs, modern `LPath` data operations plus `LatchFile`/`LatchDir`, transactional Registry APIs and samplesheets, CPU/GPU resource sizing, UI metadata and launch plans, staging and programmatic execution, Nextflow and version-aware Snakemake packaging, ready-to-use workflow discovery, automations, and the OAuth-based Latch MCP
+- **[Nextflow](../skills/nextflow/)** - Build, run, and debug Nextflow data pipelines and nf-core workflows end to end. Covers writing and testing DSL2 modules/subworkflows (processes, channels, operators, nf-test), running community pipelines (nf-core/rnaseq, nf-core/sarek), authoring samplesheets and `nextflow.config`, configuring executors and containers (Docker, Singularity/Apptainer, Conda, Wave), scaling to HPC/SLURM or cloud (AWS Batch, Google Batch, Azure, Kubernetes), and debugging failed or `-resume` runs. Use for any reproducible scientific/bioinformatics workflow work and for authoring nf-core-compliant pipelines, modules, configs, and linting
+- **[pacsomatic](../skills/pacsomatic/)** - Operator toolkit for running the nf-core/pacsomatic matched tumor-normal (somatic variant calling) workflow from BAM inputs. Validates run inputs, generates pacsomatic-compliant samplesheets, prepares reproducible Nextflow launch artifacts, runs locally or submits to schedulers (LSF/Slurm/PBS/SGE), and triages execution and startup failures. Use to prepare launch commands/scripts, perform dry-run checks, or troubleshoot pipeline and scheduler submission errors
 
 ### Microscopy & Bio-image Data
-- **OMERO Integration** - Toolkit for interacting with OMERO microscopy data management systems using Python. Provides comprehensive access to microscopy images stored in OMERO servers, including dataset and screening data retrieval, pixel data analysis, annotation and metadata management, regions of interest (ROIs) creation and analysis, batch processing, OMERO.scripts development, and OMERO.tables for structured data storage. Essential for researchers working with high-content screening data, multi-dimensional microscopy datasets, or collaborative image repositories
+- **[OMERO Integration](../skills/omero-integration/)** - Secure, scope-limited OMERO.server workflows using OMERO.py 5.22.1/BlitzGateway, the OMERO CLI, tables, annotations, ROIs, rendering, and documented OMERO.web APIs. Start with local validation or a dry run, bound every remote inventory/export, minimize sensitive output, and treat imports, annotation changes, scripts, table writes, ownership changes, and deletion as separately reviewed writes. OMERO currently supports Python 3.10-3.12
 
 ### Protocol Management & Sharing
-- **Protocols.io Integration** - Integration with protocols.io API for managing scientific protocols. Enables programmatic access to protocol discovery (search by keywords, DOI, category), protocol lifecycle management (create, update, publish with DOI), step-by-step procedure documentation, collaborative development with workspaces and discussions, file management (upload data, images, documents), experiment tracking and documentation, and data export. Supports OAuth authentication, protocol PDF generation, materials management, threaded comments, workspace permissions, and institutional protocol repositories. Essential for protocol standardization, reproducibility, lab knowledge management, and scientific collaboration
+- **[Protocols.io Integration](../skills/protocolsio-integration/)** - Read, validate, and safely export protocols.io data against the exact documented endpoint version: the maintained API mixes v3 and v4 contracts. Bundled Python 3.11+ clients are offline by default, perform only bounded official-host GET requests when `--execute` is explicit, and never execute mutations; create/update/publish operations remain non-executing plans for authorized review
 
 ## Scientific Packages
 
 ### Bioinformatics & Genomics
-- **AnnData** - Python package for handling annotated data matrices, specifically designed for single-cell genomics data. Provides efficient storage and manipulation of high-dimensional data with associated annotations (observations/cells and variables/genes). Key features include: HDF5-based h5ad file format for efficient I/O and compression, integration with pandas DataFrames for metadata, support for sparse matrices (scipy.sparse) for memory efficiency, layered data organization (X for main data matrix, obs for observation annotations, var for variable annotations, obsm/varm for multi-dimensional annotations, obsp/varp for pairwise matrices), and seamless integration with Scanpy, scvi-tools, and other single-cell analysis packages. Supports lazy loading, chunked operations, and conversion to/from other formats (CSV, HDF5, Zarr). Use cases: single-cell RNA-seq data management, multi-modal single-cell data (RNA+ATAC, CITE-seq), spatial transcriptomics, and any high-dimensional annotated data requiring efficient storage and manipulation
-- **Arboreto** - Scalable Python GRN inference (Aerts Lab) using GRNBoost2 (gradient boosting) and GENIE3 (random forest) with Dask distributed execution. Accepts pandas DataFrames, dense NumPy arrays, or sparse CSC matrices (cells × genes). Features: local multi-core and cluster scheduling, TF-restricted search, `limit` for top-N links, and `diy()` for custom regressors. Core engine for pySCENIC (cisTarget pruning and regulon activity are downstream). Use for bulk or single-cell expression matrices when inferring TF–target regulatory links from co-expression patterns
-- **BioPython** - Comprehensive Python library for computational biology and bioinformatics providing tools for sequence manipulation, database access, and biological data analysis. Key features include: sequence objects (Seq, SeqRecord, SeqIO) for DNA/RNA/protein sequences with biological alphabet validation, file format parsers (FASTA, FASTQ, GenBank, EMBL, Swiss-Prot, PDB, SAM/BAM, VCF, GFF), NCBI database access (Entrez Programming Utilities for PubMed, GenBank, BLAST, taxonomy), BLAST integration (running searches, parsing results), sequence alignment (pairwise and multiple sequence alignment with Bio.Align), phylogenetics (tree construction and manipulation with Bio.Phylo), population genetics (Hardy-Weinberg, F-statistics), protein structure analysis (PDB parsing, structure calculations), and statistical analysis tools. Supports integration with NumPy, pandas, and other scientific Python libraries. Use cases: sequence analysis, database queries, phylogenetic analysis, sequence alignment, file format conversion, and general bioinformatics workflows
-- **BioServices** - Python library providing unified programmatic access to 40+ biological web services and databases. Supports major bioinformatics resources including KEGG (pathway and compound data), UniProt (protein sequences and annotations), ChEBI (chemical entities), ChEMBL (bioactive molecules), Reactome (pathways), IntAct (protein interactions), BioModels (biological models), and many others. Features consistent API across different services, automatic result caching, error handling and retry logic, support for both REST and SOAP web services, and conversion of results to Python objects (dictionaries, lists, BioPython objects). Handles authentication, rate limiting, and API versioning. Use cases: automated data retrieval from multiple biological databases, building bioinformatics pipelines, database integration workflows, and programmatic access to biological web resources without manual web browsing
-- **Bulk RNA-seq** - End-to-end orchestrator for bulk RNA-seq differential-expression studies, chaining the repo's skills into one defensible pipeline: raw FASTQ -> QC/trimming (FastQC, fastp/Trim Galore) -> alignment and quantification (STAR, Salmon, featureCounts) -> a gene-level counts matrix -> differential expression (PyDESeq2) -> pathway/GSEA enrichment (Pathway Enrichment) -> publication figures (scientific-visualization). Offers two interchangeable upstream paths with a decision guide — nf-core/rnaseq driven through the Nextflow skill, or standalone STAR/Salmon recipes — plus the glue no single tool provides: experimental-design and QC guidance (replication, batch/confounding, strandedness, mapping-rate and PCA gates), a quant->counts bridge script (build_counts_matrix.py: aggregates Salmon quant.sf via pytximport, selects the correct STAR ReadsPerGene strand column, or parses featureCounts into a PyDESeq2-ready counts.csv plus metadata template), and a samplesheet/design validator (validate_samplesheet.py). Emphasizes reproducibility (pinned pipeline/tool/reference versions, containers) and correct statistics (integer counts, gene-ID mapping to symbols, FDR control). Use cases: going from sequencing reads to differentially expressed genes and enriched pathways, running or configuring nf-core/rnaseq, STAR/Salmon quantification, building a counts matrix for DESeq2, and scoping a complete reproducible bulk RNA-seq analysis. For single-cell RNA-seq use Scanpy instead
-- **Cellxgene Census** - Python package for querying and analyzing versioned public single-cell and spatial transcriptomics data from CZ CELLxGENE Discover. Provides access to the current stable LTS Census release with 217M+ total cells, 125M+ unique cells, and 1,845 datasets, with standardized annotations and metadata across human, mouse, marmoset, rhesus macaque, and chimpanzee. Key features include: efficient TileDB-SOMA access for scalable metadata and expression queries, AnnData and Scanpy integration, source H5AD URI/download helpers, precomputed summary counts, embeddings, spatial Census access, and TileDB-SOMA-ML guidance for PyTorch loaders. Enables subsetting datasets by organism, cell type, tissue, disease, assay, or other metadata before downloading, reducing transfer and memory requirements. Use cases: large-scale single-cell analysis, spatial transcriptomics access, cell-type discovery, cross-dataset comparisons, reference atlas construction, and exploratory analysis of public Census data
-- **deepTools** - Comprehensive suite of Python tools for exploring and visualizing next-generation sequencing (NGS) data, particularly ChIP-seq, RNA-seq, and ATAC-seq experiments. Provides command-line tools and Python API for processing BAM and bigWig files. Key features include: quality control metrics (plotFingerprint, plotCorrelation), coverage track generation (bamCoverage for creating bigWig files), matrix generation for heatmaps (computeMatrix, plotHeatmap, plotProfile), comparative analysis (multiBigwigSummary, plotPCA), and efficient handling of large files. Supports normalization methods, binning options, and various visualization outputs. Designed for high-throughput analysis workflows and publication-quality figure generation. Use cases: ChIP-seq peak visualization, RNA-seq coverage analysis, ATAC-seq signal tracks, comparative genomics, and NGS data exploration
-- **FlowIO** - Low-level Python reader and writer for Flow Cytometry Standard files, with examples verified against FlowIO 1.4.0. Reads FCS 2.0, 3.0, and 3.1 metadata and list-mode events; exposes normalized TEXT/ANALYSIS metadata, channel labels and indices, raw or gain/log/time-scaled NumPy arrays, and legacy multi-dataset files; and writes constrained FCS 3.1 single-precision output. Includes safe offset-recovery guidance, memory/privacy caveats, round-trip validation patterns, DataFrame/CSV workflows, and a metadata-first JSON inspector. FlowIO does not apply compensation, cytometry display transforms, or gating; use FlowKit for those higher-level operations
-- **gget** - Command-line tool and Python package for efficient querying of genomic databases with a simple, unified interface. Provides fast access to Ensembl (gene information and sequences), UniProt, NCBI BLAST/BLAT and viral sequence data, PDB/AlphaFold structures, CELLxGENE Census, OpenTargets, cBioPortal, COSMIC, Enrichr, Bgee, and 8cubeDB mouse specificity/expression metrics. Features include: single-command queries without complex API setup, automatic result formatting, batch query support, integration with pandas DataFrames, and support for both command-line and Python API usage. Optimized for speed and ease of use, making database queries accessible to users without extensive bioinformatics experience. Use cases: quick gene lookups, sequence retrieval, viral dataset downloads, enrichment analysis, disease/drug associations, protein structure access, and rapid database queries in bioinformatics workflows
-- **geniml** - Genomic interval machine learning toolkit providing unsupervised methods for building ML models on BED files. Key capabilities include Region2Vec (word2vec-style embeddings of genomic regions and region sets using tokenization and neural language modeling), BEDspace (joint embeddings of regions and metadata labels using StarSpace for cross-modal queries), scEmbed (Region2Vec applied to single-cell ATAC-seq data generating cell-level embeddings for clustering and annotation with scanpy integration), consensus peak building (four statistical methods CC/CCF/ML/HMM for creating reference universes from BED collections), and comprehensive utilities (BBClient for BED caching, BEDshift for genomic randomization preserving context, evaluation metrics for embedding quality, Text2BedNN for neural search backends). Part of BEDbase ecosystem. Supports Python API and CLI workflows, pre-trained models on Hugging Face, and integration with gtars for tokenization. Use cases: region similarity searches, dimension reduction of chromatin accessibility data, scATAC-seq clustering and cell-type annotation, metadata-aware genomic queries, universe construction for standardized references, and any ML task requiring genomic region feature vectors
-- **gtars** - High-performance Rust toolkit for genomic interval analysis providing specialized tools for overlap detection using IGD (Integrated Genome Database) indexing, coverage track generation (uniwig module for WIG/BigWig formats), genomic tokenization for machine learning applications (TreeTokenizer for deep learning models), reference sequence management (refget protocol compliance), fragment processing for single-cell genomics (barcode-based splitting and cluster analysis), and fragment scoring against reference datasets. Offers Python bindings with NumPy integration, command-line tools (gtars-cli), and Rust library. Key modules include: tokenizers (convert genomic regions to ML tokens), overlaprs (efficient overlap computation), uniwig (ATAC-seq/ChIP-seq/RNA-seq coverage profiles), refget (GA4GH-compliant sequence digests), bbcache (BEDbase.org integration), scoring (fragment enrichment metrics), and fragsplit (single-cell fragment manipulation). Supports parallel processing, memory-mapped files, streaming for large datasets, and serves as foundation for geniml genomic ML package. Ideal for genomic ML preprocessing, regulatory element analysis, variant annotation, chromatin accessibility profiling, and computational genomics workflows
-- **Polars-Bio** - High-performance genomic interval operations and bioinformatics file I/O on Polars DataFrames. Provides overlap, nearest, merge, coverage, complement, and subtract operations for BED/VCF/BAM/GFF intervals. Streaming and cloud-native architecture for large-scale genomic analyses, serving as a faster alternative to bioframe
-- **pysam** - Current pysam 0.24/HTSlib workflows for SAM/BAM/CRAM, VCF/BCF, FASTA/FASTQ, and tabix-indexed data. Covers coordinate and index correctness, pileup/coverage filters, CRAM 3.1 references, wrapped samtools/bcftools operations, safe header-aware writing, and bundled inspection, QC, summary, and alignment-filtering scripts
-- **PyDESeq2** - Python implementation of the DESeq2 differential gene expression analysis method for bulk RNA-seq data. Provides statistical methods for determining differential expression between experimental conditions using negative binomial generalized linear models. Key features include: size factor estimation for library size normalization, dispersion estimation and shrinkage, hypothesis testing with Wald test or likelihood ratio test, multiple testing correction (Benjamini-Hochberg FDR), results filtering and ranking, and integration with pandas DataFrames. Handles complex experimental designs, batch effects, and replicates. Produces fold-change estimates, p-values, and adjusted p-values for each gene. Use cases: identifying differentially expressed genes between conditions, RNA-seq experiment analysis, biomarker discovery, and gene expression studies requiring rigorous statistical analysis
-- **Pathway Enrichment** - Pathway and gene-set enrichment analysis on gene lists or ranked gene data, with result interpretation. Supports over-representation analysis (ORA via Enrichr/Fisher's exact/hypergeometric), preranked and standard Gene Set Enrichment Analysis (GSEA), and single-sample scoring (ssGSEA/GSVA) using gseapy and the official g:Profiler client. Covers gene-set libraries (GO Biological Process/Molecular Function/Cellular Component, KEGG, Reactome, WikiPathways, and MSigDB collections including Hallmark, C2 canonical pathways, C5 ontology, and C7 immune signatures), gene-ID mapping (Ensembl/Entrez to symbols via Biomart, g:Convert, or mygene) and organism handling, choice of the statistical background universe, multiple-testing correction (Benjamini-Hochberg FDR vs g:Profiler g:SCS vs Bonferroni), redundancy reduction (enrichment maps, leading-edge genes, term clustering), and publication-ready tables plus dotplots/bar plots/GSEA running-score plots. Includes a CLI helper (run_enrichment.py) that runs ORA or preranked GSEA end-to-end — automatically building the ranking metric from a DESeq2 results table — and writes a results table and dotplot. Cross-references PyDESeq2 and Scanpy upstream (sources of differentially expressed genes and cluster markers) and database-lookup/gget for gene-ID mapping and Reactome/KEGG/STRING APIs. Use cases: functional interpretation of differentially expressed genes, CRISPR-screen hits, and single-cell cluster markers; GO/KEGG/Reactome/WikiPathways enrichment; preranked GSEA from DESeq2 statistics; pathway activity scoring per sample or cell; and building defensible, reproducible enrichment analyses that avoid common pitfalls (gene-ID/organism mismatch, wrong background, thresholding before GSEA)
-- **Scanpy** - Comprehensive Python toolkit for single-cell RNA-seq data analysis built on AnnData (scanpy 1.12.x; Python 3.12+). Provides end-to-end workflows for preprocessing (quality control, scrublet doublet detection, normalization, log transformation), dimensionality reduction (PCA, UMAP, t-SNE), Leiden clustering, marker gene identification, pseudobulk aggregation via `get.aggregate()`, trajectory inference (PAGA, diffusion maps), and visualization. Key features include: efficient handling of large datasets using sparse matrices and experimental Dask out-of-core support, integration with scvi-tools for advanced analysis, batch correction methods (ComBat), and publication-quality plotting. Optional GPU acceleration via rapids-singlecell. Use cases: single-cell RNA-seq analysis, cell-type identification, exploratory cluster markers, pseudobulk DE workflows (with pydeseq2), trajectory analysis, and comprehensive single-cell genomics workflows
-- **scVelo** - RNA velocity analysis for estimating cell state transitions from unspliced/spliced mRNA dynamics. Infers trajectory directions, computes latent time, and identifies driver genes in single-cell RNA-seq data. Complements Scanpy/scVI-tools for trajectory inference, enabling the study of cellular differentiation dynamics and lineage decisions at single-cell resolution
-- **scvi-tools** - Probabilistic deep learning models for single-cell omics analysis. PyTorch-based framework providing variational autoencoders (VAEs) for dimensionality reduction, batch correction, differential expression, and data integration across modalities. Includes 30+ models: scVI/scANVI (RNA-seq integration and cell type annotation), totalVI/totalANVI (CITE-seq protein+RNA), MultiVI (multiome RNA+ATAC integration), PeakVI (ATAC-seq analysis), DestVI/Stereoscope/Tangram (spatial transcriptomics deconvolution), MethylVI (methylation), CytoVI (flow/mass cytometry), VeloVI (RNA velocity), contrastiveVI (perturbation studies), and Solo (doublet detection). Supports seamless integration with Scanpy/AnnData ecosystem, GPU acceleration, reference mapping (scArches), and probabilistic differential expression with uncertainty quantification
-- **scikit-bio** - Python library for bioinformatics providing data structures, algorithms, and parsers for biological sequence analysis. Built on NumPy, SciPy, and pandas. Key features include: sequence objects (DNA, RNA, protein sequences) with biological alphabet validation, sequence alignment algorithms (local, global, semiglobal), phylogenetic tree manipulation, diversity metrics (alpha diversity, beta diversity, phylogenetic diversity), distance metrics for sequences and communities, file format parsers (FASTA, FASTQ, QIIME formats, Newick), and statistical analysis tools. Provides scikit-learn compatible transformers for machine learning workflows. Supports efficient processing of large sequence datasets. Use cases: sequence analysis, microbial ecology (16S rRNA analysis), metagenomics, phylogenetic analysis, and bioinformatics research requiring sequence manipulation and diversity calculations
-- **TileDB-VCF** - High-performance C++ library with Python and CLI interfaces for efficient storage and retrieval of genomic variant-call data using TileDB multidimensional sparse array technology. Enables scalable VCF/BCF ingestion with incremental sample addition, compressed storage, parallel queries across genomic regions and samples, and export capabilities for population genomics workflows. Key features include: memory-efficient queries, cloud storage integration (S3, Azure, GCS), and CLI tools for dataset creation, sample ingestion, data export, and statistics. Supports building variant databases for large cohorts, population-scale genomics studies, and association analysis. Use cases: population genomics databases, cohort studies, variant discovery workflows, genomic data warehousing, and scaling to enterprise-level analysis with TileDB-Cloud platform
-- **Zarr** - Python library (Zarr-Python 3.x) implementing chunked, compressed N-dimensional arrays for local disk and cloud object storage (S3, GCS via fsspec). Supports Zarr format 2 and 3, `zarr.codecs` compression (Blosc, gzip, zstd), partial chunk reads, consolidated metadata, sharding, and integration with NumPy, Dask, and Xarray. Use for out-of-core arrays, cloud-native pipelines, and large scientific datasets (genomics, imaging, climate). Skill: `zarr-python`
+- **[AnnData](../skills/anndata/)** - Python package for handling annotated data matrices, specifically designed for single-cell genomics data. Provides efficient storage and manipulation of high-dimensional data with associated annotations (observations/cells and variables/genes). Key features include: HDF5-based h5ad file format for efficient I/O and compression, integration with pandas DataFrames for metadata, support for sparse matrices (scipy.sparse) for memory efficiency, layered data organization (X for main data matrix, obs for observation annotations, var for variable annotations, obsm/varm for multi-dimensional annotations, obsp/varp for pairwise matrices), and seamless integration with Scanpy, scvi-tools, and other single-cell analysis packages. Supports lazy loading, chunked operations, and conversion to/from other formats (CSV, HDF5, Zarr). Use cases: single-cell RNA-seq data management, multi-modal single-cell data (RNA+ATAC, CITE-seq), spatial transcriptomics, and any high-dimensional annotated data requiring efficient storage and manipulation
+- **[Arboreto](../skills/arboreto/)** - Scalable Python GRN inference (Aerts Lab) using GRNBoost2 (gradient boosting) and GENIE3 (random forest) with Dask distributed execution. Accepts pandas DataFrames, dense NumPy arrays, or sparse CSC matrices (cells × genes). Features: local multi-core and cluster scheduling, TF-restricted search, `limit` for top-N links, and `diy()` for custom regressors. Core engine for pySCENIC (cisTarget pruning and regulon activity are downstream). Use for bulk or single-cell expression matrices when inferring TF–target regulatory links from co-expression patterns
+- **[BioPython](../skills/biopython/)** - Comprehensive Python library for computational biology and bioinformatics providing tools for sequence manipulation, database access, and biological data analysis. Key features include: sequence objects (Seq, SeqRecord, SeqIO) for DNA/RNA/protein sequences with biological alphabet validation, file format parsers (FASTA, FASTQ, GenBank, EMBL, Swiss-Prot, PDB, SAM/BAM, VCF, GFF), NCBI database access (Entrez Programming Utilities for PubMed, GenBank, BLAST, taxonomy), BLAST integration (running searches, parsing results), sequence alignment (pairwise and multiple sequence alignment with Bio.Align), phylogenetics (tree construction and manipulation with Bio.Phylo), population genetics (Hardy-Weinberg, F-statistics), protein structure analysis (PDB parsing, structure calculations), and statistical analysis tools. Supports integration with NumPy, pandas, and other scientific Python libraries. Use cases: sequence analysis, database queries, phylogenetic analysis, sequence alignment, file format conversion, and general bioinformatics workflows
+- **[BioServices](../skills/bioservices/)** - Python library providing unified programmatic access to 40+ biological web services and databases. Supports major bioinformatics resources including KEGG (pathway and compound data), UniProt (protein sequences and annotations), ChEBI (chemical entities), ChEMBL (bioactive molecules), Reactome (pathways), IntAct (protein interactions), BioModels (biological models), and many others. Features consistent API across different services, automatic result caching, error handling and retry logic, support for both REST and SOAP web services, and conversion of results to Python objects (dictionaries, lists, BioPython objects). Handles authentication, rate limiting, and API versioning. Use cases: automated data retrieval from multiple biological databases, building bioinformatics pipelines, database integration workflows, and programmatic access to biological web resources without manual web browsing
+- **[Bulk RNA-seq](../skills/bulk-rnaseq/)** - End-to-end orchestrator for bulk RNA-seq differential-expression studies, chaining the repo's skills into one defensible pipeline: raw FASTQ -> QC/trimming (FastQC, fastp/Trim Galore) -> alignment and quantification (STAR, Salmon, featureCounts) -> a gene-level counts matrix -> differential expression (PyDESeq2) -> pathway/GSEA enrichment (Pathway Enrichment) -> publication figures (scientific-visualization). Offers two interchangeable upstream paths with a decision guide — nf-core/rnaseq driven through the Nextflow skill, or standalone STAR/Salmon recipes — plus the glue no single tool provides: experimental-design and QC guidance (replication, batch/confounding, strandedness, mapping-rate and PCA gates), a quant->counts bridge script (build_counts_matrix.py: aggregates Salmon quant.sf via pytximport, selects the correct STAR ReadsPerGene strand column, or parses featureCounts into a PyDESeq2-ready counts.csv plus metadata template), and a samplesheet/design validator (validate_samplesheet.py). Emphasizes reproducibility (pinned pipeline/tool/reference versions, containers) and correct statistics (integer counts, gene-ID mapping to symbols, FDR control). Use cases: going from sequencing reads to differentially expressed genes and enriched pathways, running or configuring nf-core/rnaseq, STAR/Salmon quantification, building a counts matrix for DESeq2, and scoping a complete reproducible bulk RNA-seq analysis. For single-cell RNA-seq use Scanpy instead
+- **[Cellxgene Census](../skills/cellxgene-census/)** - Python package for querying and analyzing versioned public single-cell and spatial transcriptomics data from CZ CELLxGENE Discover. Provides access to the current stable LTS Census release with 217M+ total cells, 125M+ unique cells, and 1,845 datasets, with standardized annotations and metadata across human, mouse, marmoset, rhesus macaque, and chimpanzee. Key features include: efficient TileDB-SOMA access for scalable metadata and expression queries, AnnData and Scanpy integration, source H5AD URI/download helpers, precomputed summary counts, embeddings, spatial Census access, and TileDB-SOMA-ML guidance for PyTorch loaders. Enables subsetting datasets by organism, cell type, tissue, disease, assay, or other metadata before downloading, reducing transfer and memory requirements. Use cases: large-scale single-cell analysis, spatial transcriptomics access, cell-type discovery, cross-dataset comparisons, reference atlas construction, and exploratory analysis of public Census data
+- **[deepTools](../skills/deeptools/)** - Comprehensive suite of Python tools for exploring and visualizing next-generation sequencing (NGS) data, particularly ChIP-seq, RNA-seq, and ATAC-seq experiments. Provides command-line tools and Python API for processing BAM and bigWig files. Key features include: quality control metrics (plotFingerprint, plotCorrelation), coverage track generation (bamCoverage for creating bigWig files), matrix generation for heatmaps (computeMatrix, plotHeatmap, plotProfile), comparative analysis (multiBigwigSummary, plotPCA), and efficient handling of large files. Supports normalization methods, binning options, and various visualization outputs. Designed for high-throughput analysis workflows and publication-quality figure generation. Use cases: ChIP-seq peak visualization, RNA-seq coverage analysis, ATAC-seq signal tracks, comparative genomics, and NGS data exploration
+- **[FlowIO](../skills/flowio/)** - Low-level Python reader and writer for Flow Cytometry Standard files, with examples verified against FlowIO 1.4.0. Reads FCS 2.0, 3.0, and 3.1 metadata and list-mode events; exposes normalized TEXT/ANALYSIS metadata, channel labels and indices, raw or gain/log/time-scaled NumPy arrays, and legacy multi-dataset files; and writes constrained FCS 3.1 single-precision output. Includes safe offset-recovery guidance, memory/privacy caveats, round-trip validation patterns, DataFrame/CSV workflows, and a metadata-first JSON inspector. FlowIO does not apply compensation, cytometry display transforms, or gating; use FlowKit for those higher-level operations
+- **[gget](../skills/gget/)** - Command-line tool and Python package for efficient querying of genomic databases with a simple, unified interface. Provides fast access to Ensembl (gene information and sequences), UniProt, NCBI BLAST/BLAT and viral sequence data, PDB/AlphaFold structures, CELLxGENE Census, OpenTargets, cBioPortal, COSMIC, Enrichr, Bgee, and 8cubeDB mouse specificity/expression metrics. Features include: single-command queries without complex API setup, automatic result formatting, batch query support, integration with pandas DataFrames, and support for both command-line and Python API usage. Optimized for speed and ease of use, making database queries accessible to users without extensive bioinformatics experience. Use cases: quick gene lookups, sequence retrieval, viral dataset downloads, enrichment analysis, disease/drug associations, protein structure access, and rapid database queries in bioinformatics workflows
+- **[Genomic Intelligence](../skills/genomic-intelligence/)** - Run six hosted transformer-DNA inference tasks from a gene symbol, genomic region, or DNA/FASTA sequence: promoter, splice-site, enhancer, chromatin-state, sequence-to-expression, and de-novo gene annotation, plus a composite gene/expression workflow. Uses a keyless capped MCP demo or keyed REST API from Python 3.10+; no local GPU is needed. Predictions are for research and development, not clinical or diagnostic decisions
+- **[geniml](../skills/geniml/)** - Audited local genomic-interval ML workflows with geniml 0.8.4 and Gtars 0.9.2: validate BED/universe contracts, plan Region2Vec or scEmbed runs, inspect model/tokenizer compatibility, and assess consensus universes. Bundled planners are Python 3.10+, dependency-free, local-only, and non-executing; remote models, caches, and training require explicit resource and network review
+- **[Gtars](../skills/gtars/)** - Local genomic interval models and set algebra, overlaps/counts, consensus/coverage, tokenization, fragment processing, and refget/BEDbase planning across Python, Rust, and the CLI. The Python binding is 0.9.2 while the Rust meta-crate/CLI are independently versioned at 0.9.0; remote constructors, pretrained tokenizers, refget, and BEDbase caching require explicit network and storage approval
+- **[OneKGPd](../skills/onekgpd/)** - Query the public high-coverage 1000 Genomes Project cohort of 3,202 GRCh38 whole genomes at individual and variant level. Supports carrier and homozygous-reference selection, per-sample regional variants, allele-frequency/functional filters, and pairwise relatedness; variant/sample queries use the public TLS endpoint, while bundled population metadata works offline without credentials
+- **[Polars-Bio](../skills/polars-bio/)** - High-performance genomic interval operations and bioinformatics file I/O on Polars DataFrames. Provides overlap, nearest, merge, coverage, complement, and subtract operations for BED/VCF/BAM/GFF intervals. Streaming and cloud-native architecture for large-scale genomic analyses, serving as a faster alternative to bioframe
+- **[pysam](../skills/pysam/)** - Current pysam 0.24/HTSlib workflows for SAM/BAM/CRAM, VCF/BCF, FASTA/FASTQ, and tabix-indexed data. Covers coordinate and index correctness, pileup/coverage filters, CRAM 3.1 references, wrapped samtools/bcftools operations, safe header-aware writing, and bundled inspection, QC, summary, and alignment-filtering scripts
+- **[PyDESeq2](../skills/pydeseq2/)** - Python implementation of the DESeq2 differential gene expression analysis method for bulk RNA-seq data. Provides statistical methods for determining differential expression between experimental conditions using negative binomial generalized linear models. Key features include: size factor estimation for library size normalization, dispersion estimation and shrinkage, hypothesis testing with Wald test or likelihood ratio test, multiple testing correction (Benjamini-Hochberg FDR), results filtering and ranking, and integration with pandas DataFrames. Handles complex experimental designs, batch effects, and replicates. Produces fold-change estimates, p-values, and adjusted p-values for each gene. Use cases: identifying differentially expressed genes between conditions, RNA-seq experiment analysis, biomarker discovery, and gene expression studies requiring rigorous statistical analysis
+- **[Pathway Enrichment](../skills/pathway-enrichment/)** - Pathway and gene-set enrichment analysis on gene lists or ranked gene data, with result interpretation. Supports over-representation analysis (ORA via Enrichr/Fisher's exact/hypergeometric), preranked and standard Gene Set Enrichment Analysis (GSEA), and single-sample scoring (ssGSEA/GSVA) using gseapy and the official g:Profiler client. Covers gene-set libraries (GO Biological Process/Molecular Function/Cellular Component, KEGG, Reactome, WikiPathways, and MSigDB collections including Hallmark, C2 canonical pathways, C5 ontology, and C7 immune signatures), gene-ID mapping (Ensembl/Entrez to symbols via Biomart, g:Convert, or mygene) and organism handling, choice of the statistical background universe, multiple-testing correction (Benjamini-Hochberg FDR vs g:Profiler g:SCS vs Bonferroni), redundancy reduction (enrichment maps, leading-edge genes, term clustering), and publication-ready tables plus dotplots/bar plots/GSEA running-score plots. Includes a CLI helper (run_enrichment.py) that runs ORA or preranked GSEA end-to-end — automatically building the ranking metric from a DESeq2 results table — and writes a results table and dotplot. Cross-references PyDESeq2 and Scanpy upstream (sources of differentially expressed genes and cluster markers) and database-lookup/gget for gene-ID mapping and Reactome/KEGG/STRING APIs. Use cases: functional interpretation of differentially expressed genes, CRISPR-screen hits, and single-cell cluster markers; GO/KEGG/Reactome/WikiPathways enrichment; preranked GSEA from DESeq2 statistics; pathway activity scoring per sample or cell; and building defensible, reproducible enrichment analyses that avoid common pitfalls (gene-ID/organism mismatch, wrong background, thresholding before GSEA)
+- **[Scanpy](../skills/scanpy/)** - Comprehensive Python toolkit for single-cell RNA-seq data analysis built on AnnData (scanpy 1.12.x; Python 3.12+). Provides end-to-end workflows for preprocessing (quality control, scrublet doublet detection, normalization, log transformation), dimensionality reduction (PCA, UMAP, t-SNE), Leiden clustering, marker gene identification, pseudobulk aggregation via `get.aggregate()`, trajectory inference (PAGA, diffusion maps), and visualization. Key features include: efficient handling of large datasets using sparse matrices and experimental Dask out-of-core support, integration with scvi-tools for advanced analysis, batch correction methods (ComBat), and publication-quality plotting. Optional GPU acceleration via rapids-singlecell. Use cases: single-cell RNA-seq analysis, cell-type identification, exploratory cluster markers, pseudobulk DE workflows (with pydeseq2), trajectory analysis, and comprehensive single-cell genomics workflows
+- **[scVelo](../skills/scvelo/)** - RNA velocity analysis for estimating cell state transitions from unspliced/spliced mRNA dynamics. Infers trajectory directions, computes latent time, and identifies driver genes in single-cell RNA-seq data. Complements Scanpy/scVI-tools for trajectory inference, enabling the study of cellular differentiation dynamics and lineage decisions at single-cell resolution
+- **[scvi-tools](../skills/scvi-tools/)** - Probabilistic deep learning models for single-cell omics analysis. PyTorch-based framework providing variational autoencoders (VAEs) for dimensionality reduction, batch correction, differential expression, and data integration across modalities. Includes 30+ models: scVI/scANVI (RNA-seq integration and cell type annotation), totalVI/totalANVI (CITE-seq protein+RNA), MultiVI (multiome RNA+ATAC integration), PeakVI (ATAC-seq analysis), DestVI/Stereoscope/Tangram (spatial transcriptomics deconvolution), MethylVI (methylation), CytoVI (flow/mass cytometry), VeloVI (RNA velocity), contrastiveVI (perturbation studies), and Solo (doublet detection). Supports seamless integration with Scanpy/AnnData ecosystem, GPU acceleration, reference mapping (scArches), and probabilistic differential expression with uncertainty quantification
+- **[scikit-bio](../skills/scikit-bio/)** - Python library for bioinformatics providing data structures, algorithms, and parsers for biological sequence analysis. Built on NumPy, SciPy, and pandas. Key features include: sequence objects (DNA, RNA, protein sequences) with biological alphabet validation, sequence alignment algorithms (local, global, semiglobal), phylogenetic tree manipulation, diversity metrics (alpha diversity, beta diversity, phylogenetic diversity), distance metrics for sequences and communities, file format parsers (FASTA, FASTQ, QIIME formats, Newick), and statistical analysis tools. Provides scikit-learn compatible transformers for machine learning workflows. Supports efficient processing of large sequence datasets. Use cases: sequence analysis, microbial ecology (16S rRNA analysis), metagenomics, phylogenetic analysis, and bioinformatics research requiring sequence manipulation and diversity calculations
+- **[TileDB-VCF](../skills/tiledbvcf/)** - High-performance C++ library with Python and CLI interfaces for efficient storage and retrieval of genomic variant-call data using TileDB multidimensional sparse array technology. Enables scalable VCF/BCF ingestion with incremental sample addition, compressed storage, parallel queries across genomic regions and samples, and export capabilities for population genomics workflows. Key features include: memory-efficient queries, cloud storage integration (S3, Azure, GCS), and CLI tools for dataset creation, sample ingestion, data export, and statistics. Supports building variant databases for large cohorts, population-scale genomics studies, and association analysis. Use cases: population genomics databases, cohort studies, variant discovery workflows, genomic data warehousing, and scaling to enterprise-level analysis with TileDB-Cloud platform
+- **[Zarr](../skills/zarr-python/)** - Python library (Zarr-Python 3.x) implementing chunked, compressed N-dimensional arrays for local disk and cloud object storage (S3, GCS via fsspec). Supports Zarr format 2 and 3, `zarr.codecs` compression (Blosc, gzip, zstd), partial chunk reads, consolidated metadata, sharding, and integration with NumPy, Dask, and Xarray. Use for out-of-core arrays, cloud-native pipelines, and large scientific datasets (genomics, imaging, climate). Skill: `zarr-python`
 
 ### Data Management & Infrastructure
-- **LaminDB** - Open-source lineage-native lakehouse for biology that makes datasets and models queryable, traceable, validated, reproducible, and FAIR. Covers LaminDB 2.x setup, artifact and collection registration, feature-based query/search, lineage tracking with `ln.track()`/`ln.finish()` and `@ln.flow()`/`@ln.step()`, schema validation for DataFrame/AnnData/SpatialData/TileDB-SOMA, and ontology-backed annotation with Bionty 2.x. Includes current guidance for projects, branches, spaces, storage backends (local, S3, GCS, S3-compatible, Hugging Face), SQLite/Postgres deployment, safe credential handling, external-data validation, and integrations with Nextflow/nf-lamin, Snakemake, Redun, W&B, MLflow, Lightning, scVI-tools, DuckDB, and Vitessce.
-- **Modal** - Serverless cloud platform for running Python code with minimal configuration, specialized for AI/ML workloads and scientific computing. Execute functions on powerful GPUs (T4, L4, A10, A100, L40S, H100, H200, B200, B200+), scale automatically from zero to thousands of containers, and pay only for compute used. Key features include: declarative container image building with uv (recommended)/pip/apt package management, automatic autoscaling with configurable limits and buffer containers, GPU acceleration with multi-GPU support (up to 8 GPUs per container, up to 1,536 GB VRAM), persistent storage via Volumes (v1 and v2) for model weights and datasets, secret management for API keys and credentials, scheduled jobs with cron expressions, web endpoints for deploying serverless APIs (FastAPI, ASGI, WSGI, WebSockets), parallel execution with `.map()` for batch processing, input concurrency and dynamic batching for I/O-bound workloads, Sandboxes for isolated execution of untrusted or agent-generated code with network (CIDR) restrictions, and resource configuration (CPU cores, memory, ephemeral disk up to 3 TiB). Supports custom Docker images, Micromamba/Conda environments, integration with Hugging Face/Weights & Biases, and distributed multi-GPU training. Free tier includes $30/month credits. Use cases: ML model deployment and inference (LLMs, image generation, speech, embeddings), GPU-accelerated training and fine-tuning, batch processing large datasets in parallel, scheduled compute-intensive jobs, serverless API deployment with autoscaling, protein folding and computational biology, scientific computing requiring distributed compute or specialized hardware, and data pipeline automation
-- **Optimize for GPU** - GPU-accelerate Python code using the NVIDIA RAPIDS ecosystem and related libraries for 10x–1000x speedups on suitable workloads. Covers 12 GPU libraries with decision framework for choosing the right tool: CuPy (drop-in NumPy/SciPy replacement for array operations, FFT, linear algebra), Numba CUDA (custom GPU kernels with fine-grained thread/block/shared-memory control), Warp (JIT-compiled simulation kernels with built-in spatial types for physics, mesh ray casting, differentiable programming, and robotics), cuDF (drop-in pandas replacement for dataframe ETL, groupby, joins), cuML (drop-in scikit-learn replacement for classification, regression, clustering, dimensionality reduction, preprocessing), cuGraph (drop-in NetworkX replacement for PageRank, centrality, community detection, shortest paths), KvikIO (GPUDirect Storage for high-performance file IO bypassing CPU memory, S3/HTTP direct-to-GPU reads, Zarr GPU backend), cuxfilter (GPU-accelerated interactive cross-filtering dashboards with Bokeh, Datashader, and Deck.gl), cuCIM (drop-in scikit-image replacement for image filtering, morphology, segmentation, plus fast whole-slide image reading for digital pathology), cuVS (GPU-accelerated vector/similarity search with CAGRA, IVF-Flat, IVF-PQ for RAG and recommender systems), cuSpatial (GPU-accelerated GeoPandas replacement for spatial joins, point-in-polygon, trajectory analysis), and RAFT/pylibraft (sparse eigensolvers, device memory management, multi-GPU communication). All libraries interoperate via CUDA Array Interface for zero-copy data sharing. Includes optimization workflow (profile first, assess GPU suitability, start with drop-in replacements, minimize host-device transfers), code transformation patterns for each library, memory management principles, and common pitfalls. Use cases: accelerating NumPy/pandas/scikit-learn/NetworkX/scikit-image/GeoPandas/Faiss workloads, physics simulation, differentiable rendering, particle systems, vector search for RAG pipelines, GPUDirect Storage file IO, interactive data exploration dashboards, geospatial analysis, medical imaging, and sparse eigenvalue problems
+- **[LaminDB](../skills/lamindb/)** - Open-source lineage-native lakehouse for biology that makes datasets and models queryable, traceable, validated, reproducible, and FAIR. Covers LaminDB 2.x setup, artifact and collection registration, feature-based query/search, lineage tracking with `ln.track()`/`ln.finish()` and `@ln.flow()`/`@ln.step()`, schema validation for DataFrame/AnnData/SpatialData/TileDB-SOMA, and ontology-backed annotation with Bionty 2.x. Includes current guidance for projects, branches, spaces, storage backends (local, S3, GCS, S3-compatible, Hugging Face), SQLite/Postgres deployment, safe credential handling, external-data validation, and integrations with Nextflow/nf-lamin, Snakemake, Redun, W&B, MLflow, Lightning, scVI-tools, DuckDB, and Vitessce.
+- **[Modal](../skills/modal/)** - Serverless cloud platform for running Python code with minimal configuration, specialized for AI/ML workloads and scientific computing. Execute functions on powerful GPUs (T4, L4, A10, A100, L40S, H100, H200, B200, B200+), scale automatically from zero to thousands of containers, and pay only for compute used. Key features include: declarative container image building with uv (recommended)/pip/apt package management, automatic autoscaling with configurable limits and buffer containers, GPU acceleration with multi-GPU support (up to 8 GPUs per container, up to 1,536 GB VRAM), persistent storage via Volumes (v1 and v2) for model weights and datasets, secret management for API keys and credentials, scheduled jobs with cron expressions, web endpoints for deploying serverless APIs (FastAPI, ASGI, WSGI, WebSockets), parallel execution with `.map()` for batch processing, input concurrency and dynamic batching for I/O-bound workloads, Sandboxes for isolated execution of untrusted or agent-generated code with network (CIDR) restrictions, and resource configuration (CPU cores, memory, ephemeral disk up to 3 TiB). Supports custom Docker images, Micromamba/Conda environments, integration with Hugging Face/Weights & Biases, and distributed multi-GPU training. Free tier includes $30/month credits. Use cases: ML model deployment and inference (LLMs, image generation, speech, embeddings), GPU-accelerated training and fine-tuning, batch processing large datasets in parallel, scheduled compute-intensive jobs, serverless API deployment with autoscaling, protein folding and computational biology, scientific computing requiring distributed compute or specialized hardware, and data pipeline automation
+- **[Optimize for GPU](../skills/optimize-for-gpu/)** - GPU-accelerate Python code using the NVIDIA RAPIDS ecosystem and related libraries for 10x–1000x speedups on suitable workloads. Covers 12 GPU libraries with decision framework for choosing the right tool: CuPy (drop-in NumPy/SciPy replacement for array operations, FFT, linear algebra), Numba CUDA (custom GPU kernels with fine-grained thread/block/shared-memory control), Warp (JIT-compiled simulation kernels with built-in spatial types for physics, mesh ray casting, differentiable programming, and robotics), cuDF (drop-in pandas replacement for dataframe ETL, groupby, joins), cuML (drop-in scikit-learn replacement for classification, regression, clustering, dimensionality reduction, preprocessing), cuGraph (drop-in NetworkX replacement for PageRank, centrality, community detection, shortest paths), KvikIO (GPUDirect Storage for high-performance file IO bypassing CPU memory, S3/HTTP direct-to-GPU reads, Zarr GPU backend), cuxfilter (GPU-accelerated interactive cross-filtering dashboards with Bokeh, Datashader, and Deck.gl), cuCIM (drop-in scikit-image replacement for image filtering, morphology, segmentation, plus fast whole-slide image reading for digital pathology), cuVS (GPU-accelerated vector/similarity search with CAGRA, IVF-Flat, IVF-PQ for RAG and recommender systems), cuSpatial (GPU-accelerated GeoPandas replacement for spatial joins, point-in-polygon, trajectory analysis), and RAFT/pylibraft (sparse eigensolvers, device memory management, multi-GPU communication). All libraries interoperate via CUDA Array Interface for zero-copy data sharing. Includes optimization workflow (profile first, assess GPU suitability, start with drop-in replacements, minimize host-device transfers), code transformation patterns for each library, memory management principles, and common pitfalls. Use cases: accelerating NumPy/pandas/scikit-learn/NetworkX/scikit-image/GeoPandas/Faiss workloads, physics simulation, differentiable rendering, particle systems, vector search for RAG pipelines, GPUDirect Storage file IO, interactive data exploration dashboards, geospatial analysis, medical imaging, and sparse eigenvalue problems
 
 ### Cheminformatics & Drug Discovery
-- **Datamol** - Python library for molecular manipulation and featurization built on RDKit with enhanced workflows and performance optimizations. Provides utilities for molecular I/O (reading/writing SMILES, SDF, MOL files), molecular standardization and sanitization, molecular transformations (tautomer enumeration, stereoisomer generation), molecular featurization (descriptors, fingerprints, graph representations), parallel processing for large datasets, and integration with machine learning pipelines. Features include: optimized RDKit operations, caching for repeated computations, molecular filtering and preprocessing, and seamless integration with pandas DataFrames. Designed for drug discovery and cheminformatics workflows requiring efficient processing of large compound libraries. Use cases: molecular preprocessing for ML models, compound library management, molecular similarity searches, and cheminformatics data pipelines
-- **DeepChem** - Molecular machine learning library (PyPI 2.8.0; Python 3.7–3.11) with TensorFlow, PyTorch, and JAX backends. Graph neural networks (GCN, GAT, MPNN, AttentiveFP), 30+ MoleculeNet benchmarks, diverse featurizers (fingerprints, descriptors, graphs), and pretrained models (ChemBERTa, GROVER, MolFormer). Install backend-specific extras (`deepchem[torch]`, `[tensorflow]`, `[jax]`). Use for ADMET/toxicity prediction, materials properties, and transfer learning on small datasets
-- **DiffDock** - Diffusion-based molecular docking method for predicting protein-ligand binding poses with DiffDock-L confidence scores. Uses diffusion models to generate diverse, high-quality poses without exhaustive search. Key features include: fast GPU inference compared to traditional docking methods, generation of multiple diverse poses, confidence scoring for pose quality, batch CSV input for screening campaigns, and support for protein PDB files or ESMFold sequence inputs with SMILES/SDF/MOL2 ligands. DiffDock does not directly predict binding affinity; combine generated poses with GNINA, MM/GBSA, or free-energy workflows when affinity ranking is required. Use cases: virtual screening, lead optimization, binding pose prediction, structure-based drug design, and initial pose generation for downstream refinement
-- **MedChem** - Python library for molecular filtering and prioritization in drug discovery (datamol-io). Apply literature-derived drug-likeness rules (Lipinski, Veber, CNS, lead-like), named structural alert catalogs (PAINS, Brenk, NIBR), complexity thresholds against ZINC-derived percentiles, chemical group detection, and a custom query language for multi-criteria filtering. Returns pandas DataFrames from rule/structural filters and boolean masks from the functional API. Integrates with datamol/RDKit and supports parallel processing. Use cases: compound library triage, hit-to-lead filtering, PAINS deprioritization, screening-deck curation, and lead optimization workflows
-- **Molfeat** - Comprehensive Python library providing 100+ molecular featurizers for converting molecules into numerical representations suitable for machine learning. Includes molecular fingerprints (ECFP, MACCS, RDKit, Pharmacophore), molecular descriptors (2D/3D descriptors, constitutional, topological, electronic), graph-based representations (molecular graphs, line graphs), and pre-trained models (MolBERT, ChemBERTa, Uni-Mol embeddings). Features unified API across different featurizer types, caching for performance, parallel processing, and integration with popular ML frameworks (scikit-learn, PyTorch, TensorFlow). Supports both traditional cheminformatics descriptors and modern learned representations. Use cases: molecular property prediction, virtual screening, molecular similarity searches, and preparing molecular data for machine learning models
-- **PyTDC** - Python library providing access to Therapeutics Data Commons (TDC), a collection of curated datasets and benchmarks for drug discovery and development. Includes datasets for ADMET prediction (absorption, distribution, metabolism, excretion, toxicity), drug-target interactions, drug-drug interactions, drug response prediction, molecular generation, and retrosynthesis. Features standardized data formats, data loaders with automatic preprocessing, benchmark tasks with evaluation metrics, leaderboards for model comparison, and integration with popular ML frameworks. Provides both single-molecule and drug-pair datasets, covering various stages of drug discovery from target identification to clinical outcomes. Use cases: benchmarking ML models for drug discovery, ADMET prediction model development, drug-target interaction prediction, and drug discovery research
-- **RDKit** - Open-source cheminformatics toolkit for molecular informatics and drug discovery. Provides comprehensive functionality for molecular I/O (reading/writing SMILES, SDF, MOL, PDB files), molecular descriptors (200+ 2D and 3D descriptors), molecular fingerprints (Morgan, RDKit, MACCS, topological torsions), SMARTS pattern matching for substructure searches, molecular alignment and 3D coordinate generation, pharmacophore perception, reaction handling, and molecular drawing. Features high-performance C++ core with Python bindings, support for large molecule sets, and extensive documentation. Widely used in pharmaceutical industry and academic research. Use cases: molecular property calculation, virtual screening, molecular similarity searches, substructure matching, molecular visualization, and general cheminformatics workflows
-- **Rowan** - Cloud-based quantum chemistry platform with Python API for computational chemistry workflows. Provides access to 45+ chemistry calculations including pKa prediction, redox potentials, solubility, conformer searching, geometry optimization, protein-ligand docking (AutoDock Vina), and AI-powered protein cofolding (Chai-1, Boltz-1/2). Supports DFT, semiempirical (GFN-xTB), and neural network potential methods (AIMNet2, Egret). Key features include: automatic cloud resource allocation, unified API for diverse computational methods, RDKit-native interface for seamless cheminformatics integration, workflow organization with folders and projects, batch processing, and web interface for visualization. Requires API key from labs.rowansci.com. Use cases: molecular property prediction, structure-based drug design, virtual screening campaigns, protein-ligand binding prediction, conformational analysis, and automated computational chemistry pipelines
-- **TorchDrug** - PyTorch-based machine learning platform for drug discovery with 40+ datasets, 20+ GNN models for molecular property prediction, protein modeling, knowledge graph reasoning, molecular generation, and retrosynthesis planning
+- **[Datamol](../skills/datamol/)** - Python library for molecular manipulation and featurization built on RDKit with enhanced workflows and performance optimizations. Provides utilities for molecular I/O (reading/writing SMILES, SDF, MOL files), molecular standardization and sanitization, molecular transformations (tautomer enumeration, stereoisomer generation), molecular featurization (descriptors, fingerprints, graph representations), parallel processing for large datasets, and integration with machine learning pipelines. Features include: optimized RDKit operations, caching for repeated computations, molecular filtering and preprocessing, and seamless integration with pandas DataFrames. Designed for drug discovery and cheminformatics workflows requiring efficient processing of large compound libraries. Use cases: molecular preprocessing for ML models, compound library management, molecular similarity searches, and cheminformatics data pipelines
+- **[DeepChem](../skills/deepchem/)** - Molecular machine learning library (PyPI 2.8.0; Python 3.7–3.11) with TensorFlow, PyTorch, and JAX backends. Graph neural networks (GCN, GAT, MPNN, AttentiveFP), 30+ MoleculeNet benchmarks, diverse featurizers (fingerprints, descriptors, graphs), and pretrained models (ChemBERTa, GROVER, MolFormer). Install backend-specific extras (`deepchem[torch]`, `[tensorflow]`, `[jax]`). Use for ADMET/toxicity prediction, materials properties, and transfer learning on small datasets
+- **[DiffDock](../skills/diffdock/)** - Diffusion-based molecular docking method for predicting protein-ligand binding poses with DiffDock-L confidence scores. Uses diffusion models to generate diverse, high-quality poses without exhaustive search. Key features include: fast GPU inference compared to traditional docking methods, generation of multiple diverse poses, confidence scoring for pose quality, batch CSV input for screening campaigns, and support for protein PDB files or ESMFold sequence inputs with SMILES/SDF/MOL2 ligands. DiffDock does not directly predict binding affinity; combine generated poses with GNINA, MM/GBSA, or free-energy workflows when affinity ranking is required. Use cases: virtual screening, lead optimization, binding pose prediction, structure-based drug design, and initial pose generation for downstream refinement
+- **[MedChem](../skills/medchem/)** - Python library for molecular filtering and prioritization in drug discovery (datamol-io). Apply literature-derived drug-likeness rules (Lipinski, Veber, CNS, lead-like), named structural alert catalogs (PAINS, Brenk, NIBR), complexity thresholds against ZINC-derived percentiles, chemical group detection, and a custom query language for multi-criteria filtering. Returns pandas DataFrames from rule/structural filters and boolean masks from the functional API. Integrates with datamol/RDKit and supports parallel processing. Use cases: compound library triage, hit-to-lead filtering, PAINS deprioritization, screening-deck curation, and lead optimization workflows
+- **[Molfeat](../skills/molfeat/)** - Comprehensive Python library providing 100+ molecular featurizers for converting molecules into numerical representations suitable for machine learning. Includes molecular fingerprints (ECFP, MACCS, RDKit, Pharmacophore), molecular descriptors (2D/3D descriptors, constitutional, topological, electronic), graph-based representations (molecular graphs, line graphs), and pre-trained models (MolBERT, ChemBERTa, Uni-Mol embeddings). Features unified API across different featurizer types, caching for performance, parallel processing, and integration with popular ML frameworks (scikit-learn, PyTorch, TensorFlow). Supports both traditional cheminformatics descriptors and modern learned representations. Use cases: molecular property prediction, virtual screening, molecular similarity searches, and preparing molecular data for machine learning models
+- **[PyTDC](../skills/pytdc/)** - Use Therapeutics Data Commons through PyTDC 1.1.15 for registry discovery, approved dataset loading, task-aware splits, evaluator metrics, benchmark groups, and bounded molecular-oracle workflows. The verified environment is CPython 3.11 with setuptools 80.9.0 because of legacy `pkg_resources`; dataset, checkpoint, benchmark, and remote-oracle network/storage effects require review and explicit approval
+- **[RDKit](../skills/rdkit/)** - Open-source cheminformatics toolkit for molecular informatics and drug discovery. Provides comprehensive functionality for molecular I/O (reading/writing SMILES, SDF, MOL, PDB files), molecular descriptors (200+ 2D and 3D descriptors), molecular fingerprints (Morgan, RDKit, MACCS, topological torsions), SMARTS pattern matching for substructure searches, molecular alignment and 3D coordinate generation, pharmacophore perception, reaction handling, and molecular drawing. Features high-performance C++ core with Python bindings, support for large molecule sets, and extensive documentation. Widely used in pharmaceutical industry and academic research. Use cases: molecular property calculation, virtual screening, molecular similarity searches, substructure matching, molecular visualization, and general cheminformatics workflows
+- **[Rowan](../skills/rowan/)** - Cloud-based quantum chemistry platform with Python API for computational chemistry workflows. Provides access to 45+ chemistry calculations including pKa prediction, redox potentials, solubility, conformer searching, geometry optimization, protein-ligand docking (AutoDock Vina), and AI-powered protein cofolding (Chai-1, Boltz-1/2). Supports DFT, semiempirical (GFN-xTB), and neural network potential methods (AIMNet2, Egret). Key features include: automatic cloud resource allocation, unified API for diverse computational methods, RDKit-native interface for seamless cheminformatics integration, workflow organization with folders and projects, batch processing, and web interface for visualization. Requires API key from labs.rowansci.com. Use cases: molecular property prediction, structure-based drug design, virtual screening campaigns, protein-ligand binding prediction, conformational analysis, and automated computational chemistry pipelines
+- **[TorchDrug](../skills/torchdrug/)** - PyTorch-based machine learning platform for drug discovery with 40+ datasets, 20+ GNN models for molecular property prediction, protein modeling, knowledge graph reasoning, molecular generation, and retrosynthesis planning
 
 ### Proteomics & Mass Spectrometry
-- **matchms** - Reproducible MS/MS processing and spectral-library search with metadata/peak filtering, cosine and exact/greedy modified-cosine scoring, fast BLINK and Flash modes, structured sparse scores, spectral networks, and MGF/MSP/mzML/mzXML/JSON/mzSpecLib/USI workflows
-- **pyOpenMS** - Comprehensive mass spectrometry data analysis for proteomics and metabolomics (LC-MS/MS processing, peptide identification, feature detection, quantification, chemical calculations, and integration with search engines like Comet, Mascot, MSGF+)
+- **[matchms](../skills/matchms/)** - Reproducible MS/MS processing and spectral-library search with metadata/peak filtering, cosine and exact/greedy modified-cosine scoring, fast BLINK and Flash modes, structured sparse scores, spectral networks, and MGF/MSP/mzML/mzXML/JSON/mzSpecLib/USI workflows
+- **[pyOpenMS](../skills/pyopenms/)** - Comprehensive mass spectrometry data analysis for proteomics and metabolomics (LC-MS/MS processing, peptide identification, feature detection, quantification, chemical calculations, and integration with search engines like Comet, Mascot, MSGF+)
 
 ### Medical Imaging & Digital Pathology
-- **histolab** - Digital pathology toolkit for whole slide image (WSI) processing and analysis. Provides automated tissue detection, tile extraction for deep learning pipelines, and preprocessing for gigapixel histopathology images. Key features include: multi-format WSI support (SVS, TIFF, NDPI), three tile extraction strategies (RandomTiler for sampling, GridTiler for complete coverage, ScoreTiler for quality-driven selection), automated tissue masks with customizable filters, built-in scorers (NucleiScorer, CellularityScorer), Macenko and Reinhard stain normalization (0.6.0+), pyramidal image handling, visualization tools (thumbnails, mask overlays, tile previews), and H&E stain decomposition. Supports multiple tissue sections, artifact removal, pen annotation exclusion, and reproducible extraction with seeding. Requires Python 3.8–3.11, OpenSlide, and Linux or macOS. Use cases: creating training datasets for computational pathology, extracting informative tiles for tumor classification, whole-slide tissue characterization, quality assessment of histology samples, automated nuclei density analysis, and preprocessing for digital pathology deep learning workflows
-- **PathML** - Comprehensive computational pathology toolkit for whole slide image analysis, tissue segmentation, and machine learning on pathology data. Provides end-to-end workflows for digital pathology research including data loading, preprocessing, feature extraction, and model deployment
-- **pydicom** - Pure Python package for working with DICOM (Digital Imaging and Communications in Medicine) files. Provides comprehensive support for reading, writing, and manipulating medical imaging data from CT, MRI, X-ray, ultrasound, PET scans and other modalities. Key features include: pixel data extraction and manipulation with automatic decompression (JPEG/JPEG 2000/RLE), metadata access and modification with 1000+ standardized DICOM tags, image format conversion (PNG/JPEG/TIFF), anonymization tools for removing Protected Health Information (PHI), windowing and display transformations (VOI LUT application), multi-frame and 3D volume processing, DICOM sequence handling, and support for multiple transfer syntaxes. Use cases: medical image analysis, PACS system integration, radiology workflows, research data processing, DICOM anonymization, format conversion, image preprocessing for machine learning, multi-slice volume reconstruction, and clinical imaging pipelines
+- **[histolab](../skills/histolab/)** - Digital pathology toolkit for whole slide image (WSI) processing and analysis. Provides automated tissue detection, tile extraction for deep learning pipelines, and preprocessing for gigapixel histopathology images. Key features include: multi-format WSI support (SVS, TIFF, NDPI), three tile extraction strategies (RandomTiler for sampling, GridTiler for complete coverage, ScoreTiler for quality-driven selection), automated tissue masks with customizable filters, built-in scorers (NucleiScorer, CellularityScorer), Macenko and Reinhard stain normalization (0.6.0+), pyramidal image handling, visualization tools (thumbnails, mask overlays, tile previews), and H&E stain decomposition. Supports multiple tissue sections, artifact removal, pen annotation exclusion, and reproducible extraction with seeding. Requires Python 3.8–3.11, OpenSlide, and Linux or macOS. Use cases: creating training datasets for computational pathology, extracting informative tiles for tumor classification, whole-slide tissue characterization, quality assessment of histology samples, automated nuclei density analysis, and preprocessing for digital pathology deep learning workflows
+- **[PathML](../skills/pathml/)** - Local, research-only computational pathology with PathML 3.0.5 on Python 3.10-3.12: load and tile slides, build preprocessing/QC pipelines, manage h5path data, quantify multiplex images, construct spatial graphs, and plan bounded inference. It is beta research software, not a diagnostic system or medical device; use authorized de-identified data, patient-level splits, and approved encrypted local storage
+- **[pydicom](../skills/pydicom/)** - Read, inspect, write, transform, and preflight authorized local DICOM datasets with pydicom 3.0.2, including transfer syntaxes, compression plug-ins, frames, private elements, JSON, and bounded de-identification review. Metadata, private tags, overlays, filenames, and pixels may contain PHI; default to allowlisted aggregate output. pydicom is not a diagnostic viewer, and tag removal alone does not establish DICOM PS3.15, HIPAA, GDPR, or other compliance
 
 ### Healthcare AI & Clinical Machine Learning
-- **NeuroKit2** - Comprehensive biosignal processing toolkit for analyzing physiological data including ECG, EEG, EDA, RSP, PPG, EMG, and EOG signals. Use this skill when processing cardiovascular signals, brain activity, electrodermal responses, respiratory patterns, muscle activity, or eye movements. Key features include: automated signal processing pipelines (cleaning, peak detection, delineation, quality assessment), heart rate variability analysis across time/frequency/nonlinear domains (SDNN, RMSSD, LF/HF, DFA, entropy measures), EEG analysis (frequency band power, microstates, source localization), autonomic nervous system assessment (sympathetic indices, respiratory sinus arrhythmia), comprehensive complexity measures (25+ entropy types, 15+ fractal dimensions, Lyapunov exponents), event-related and interval-related analysis modes, epoch creation and averaging for stimulus-locked responses, multi-signal integration with unified workflows, and extensive signal processing utilities (filtering, decomposition, peak correction, spectral analysis). Includes modular reference documentation across 12 specialized domains. Use cases: heart rate variability for cardiovascular health assessment, EEG microstates for consciousness studies, electrodermal activity for emotion research, respiratory variability analysis, psychophysiology experiments, affective computing, stress monitoring, sleep staging, autonomic dysfunction assessment, biofeedback applications, and multi-modal physiological signal integration for comprehensive human state monitoring
-- **PyHealth** - Comprehensive healthcare AI toolkit for developing, testing, and deploying machine learning models with clinical data. Provides specialized tools for electronic health records (EHR), physiological signals, medical imaging, and clinical text analysis. Key features include: 10+ healthcare datasets (MIMIC-III/IV, eICU, OMOP, sleep EEG, COVID-19 CXR), 20+ predefined clinical prediction tasks (mortality, hospital readmission, length of stay, drug recommendation, sleep staging, EEG analysis), 33+ models (Logistic Regression, MLP, CNN, RNN, Transformer, GNN, plus healthcare-specific models like RETAIN, SafeDrug, GAMENet, StageNet), comprehensive data processing (sequence processors, signal processors, medical code translation between ICD-9/10, NDC, RxNorm, ATC systems), training/evaluation utilities (Trainer class, fairness metrics, calibration, uncertainty quantification), and interpretability tools (attention visualization, SHAP, ChEFER). 3x faster than pandas for healthcare data processing. Use cases: ICU mortality prediction, hospital readmission risk assessment, safe medication recommendation with drug-drug interaction constraints, sleep disorder diagnosis from EEG signals, medical code standardization and translation, clinical text to ICD coding, length of stay estimation, and any clinical ML application requiring interpretability, fairness assessment, and calibrated predictions for healthcare deployment
+- **[PyHealth](../skills/pyhealth/)** - Healthcare-ML research toolkit for EHR, signal, imaging, and text datasets, with task construction, code mapping, data processors, classical/deep models, training, calibration, fairness, uncertainty, and interpretability utilities. Use these capabilities for authorized retrospective model development and evaluation; do not turn example tasks or predictions into patient-specific diagnosis, medication recommendations, alarms, or deployment decisions without independent clinical, technical, privacy, regulatory, and institutional validation
 
 ### Clinical Documentation & Decision Support
-- **Clinical Decision Support** - Generate professional clinical decision support (CDS) documents for pharmaceutical and clinical research settings. Includes patient cohort analyses (biomarker-stratified with outcomes) and treatment recommendation reports (evidence-based guidelines with decision algorithms). Features GRADE evidence grading, statistical analysis (hazard ratios, survival curves, waterfall plots), biomarker integration (genomic alterations, gene expression signatures, IHC markers), and regulatory compliance. Use cases: pharmaceutical cohort reporting, clinical guideline development, comparative effectiveness analyses, treatment algorithm creation, and evidence synthesis for drug development
-- **Clinical Reports** - Write comprehensive clinical reports following established guidelines and standards. Covers case reports (CARE guidelines), diagnostic reports (radiology, pathology, laboratory), clinical trial reports (ICH-E3, SAE, CSR), and patient documentation (SOAP notes, H&P, discharge summaries). Includes templates, regulatory compliance (HIPAA, FDA, ICH-GCP), and validation tools. Use cases: journal case reports, diagnostic findings documentation, clinical trial reporting, patient progress notes, and regulatory submissions
-- **Treatment Plans** - Generate concise (3-4 page), focused medical treatment plans in LaTeX/PDF format for all clinical specialties. Supports general medical treatment, rehabilitation therapy, mental health care, chronic disease management, perioperative care, and pain management. Features SMART goal frameworks, evidence-based interventions, HIPAA compliance, and professional formatting. Use cases: individualized patient care plans, rehabilitation programs, psychiatric treatment plans, surgical care pathways, and pain management protocols
+- **[Clinical Decision Support](../skills/clinical-decision-support/)** - Version 2.0 prepares local, research-only evaluation, evidence-profile, aggregate cohort, survival, biomarker/model, privacy, and governance artifacts from synthetic or aggregate data. Bundled scripts use only Python 3.11+ standard library and make no network, credential, model, or image calls. The skill never diagnoses, recommends or changes treatment, calculates doses, triages, alerts, operates at point of care, or establishes regulatory/privacy compliance
+- **[Clinical Reports](../skills/clinical-reports/)** - Version 2.0 creates safety-bounded draft structures, aggregate tables, source-fact manifests, and deterministic checks for clinical case, diagnostic, trial, safety, and aggregate research reports. Optional Python 3.11+ scripts are dependency-free and local-only. Inputs must be synthetic, de-identified, or aggregate and every fact verified; outputs remain visibly marked drafts for qualified review, never diagnosis, treatment advice, source-record amendment, filing, transmission, or submission
+- **[Treatment Plans](../skills/treatment-plans/)** - Version 2.0 only formats and structurally validates local JSON documentation of decisions already supplied and verified by authorized licensed professionals; bundled Python 3.11+ tools use no network, models, images, credentials, or third-party packages. It supports source traceability, clinician-authored intervention records, goals/checkpoints, shared-decision records, reconciliation handoffs, and release gates, but never selects, compares, or modifies therapy, dose, monitoring, eligibility, or urgency
 
 ### Neuroscience & Electrophysiology
-- **BIDS** - Brain Imaging Data Structure (BIDS) standard for organizing and describing neuroscience and biomedical research datasets. While originating for MRI, BIDS now covers 11 modalities: imaging (MRI structural/functional/diffusion/perfusion, PET, microscopy), electrophysiology (EEG, MEG, iEEG, EMG), and other data (NIRS, motion capture, behavioral, MR spectroscopy), with active BEPs extending to microelectrode electrophysiology (Neuropixels), stimuli, and more. Covers the BIDS directory hierarchy, file naming conventions with entities (subject, session, task, acquisition, run, etc.), and JSON sidecar metadata. Key features include: dataset creation and validation workflows, querying BIDS datasets with PyBIDS (BIDSLayout), DICOM-to-BIDS conversion using HeuDiConv (ReproIn turnkey, map-into-reproin, and custom heuristic modes), dcm2bids (config-file-based), and BIDScoin (GUI-based), metadata inheritance and sidecar management, events files for task fMRI, participants and scans TSV files, BIDS derivatives conventions for preprocessed data and analysis outputs, BIDS-Apps interface (fMRIPrep, MRIQC, QSIPrep), machine-readable BIDS schema (bids_schema.json) and BEP listing (beps.yml) with update script, and .bidsignore configuration. Includes detailed reference documentation for the complete BIDS specification entity table (35 entities in schema ordering), required and recommended metadata fields for every modality, standard template spaces, and conversion tool workflows with examples. Use cases: organizing neuroscience data for sharing and analysis, validating BIDS compliance before repository submission (OpenNeuro, DANDI), converting DICOM scanner data to BIDS format, creating BIDS-compliant derivatives, querying datasets programmatically, and preparing data for BIDS-Apps processing pipelines
-- **Neuropixels-Analysis** - Comprehensive toolkit for analyzing Neuropixels high-density neural recordings using SpikeInterface, Allen Institute, and International Brain Laboratory (IBL) best practices. Supports the full workflow from raw data to publication-ready curated units. Key features include: data loading from SpikeGLX, Open Ephys, and NWB formats, preprocessing pipelines (highpass filtering, phase shift correction for Neuropixels 1.0, bad channel detection, common average referencing), motion/drift estimation and correction (nonrigid_fast_and_accurate, nonrigid_accurate, and DREDge presets), spike sorting integration (Kilosort4 GPU, SpykingCircus2, Tridesclous2, Mountainsort5 CPU), comprehensive postprocessing (waveform extraction, template computation, spike amplitudes, correlograms, unit locations), quality metrics computation (SNR, ISI violations, presence ratio, amplitude cutoff, drift metrics), automated curation using Allen Institute and IBL criteria with configurable thresholds, model-based curation with pretrained UnitRefine classifiers (noise/neural and SUA/MUA) loaded from Hugging Face via spikeinterface.curation, AI-assisted visual curation for uncertain units using vision-language models, and export to Phy for manual review or NWB for sharing. Supports Neuropixels 1.0 (960 electrodes, 384 channels) and Neuropixels 2.0 (single and 4-shank configurations). Use cases: extracellular electrophysiology analysis, spike sorting from silicon probes, neural population recordings, systems neuroscience research, unit quality assessment, publication-ready neural data processing, and integration of AI-assisted curation for borderline units
+- **[BIDS](../skills/bids/)** - Brain Imaging Data Structure (BIDS) standard for organizing and describing neuroscience and biomedical research datasets. While originating for MRI, BIDS now covers 11 modalities: imaging (MRI structural/functional/diffusion/perfusion, PET, microscopy), electrophysiology (EEG, MEG, iEEG, EMG), and other data (NIRS, motion capture, behavioral, MR spectroscopy), with active BEPs extending to microelectrode electrophysiology (Neuropixels), stimuli, and more. Covers the BIDS directory hierarchy, file naming conventions with entities (subject, session, task, acquisition, run, etc.), and JSON sidecar metadata. Key features include: dataset creation and validation workflows, querying BIDS datasets with PyBIDS (BIDSLayout), DICOM-to-BIDS conversion using HeuDiConv (ReproIn turnkey, map-into-reproin, and custom heuristic modes), dcm2bids (config-file-based), and BIDScoin (GUI-based), metadata inheritance and sidecar management, events files for task fMRI, participants and scans TSV files, BIDS derivatives conventions for preprocessed data and analysis outputs, BIDS-Apps interface (fMRIPrep, MRIQC, QSIPrep), machine-readable BIDS schema (bids_schema.json) and BEP listing (beps.yml) with update script, and .bidsignore configuration. Includes detailed reference documentation for the complete BIDS specification entity table (35 entities in schema ordering), required and recommended metadata fields for every modality, standard template spaces, and conversion tool workflows with examples. Use cases: organizing neuroscience data for sharing and analysis, validating BIDS compliance before repository submission (OpenNeuro, DANDI), converting DICOM scanner data to BIDS format, creating BIDS-compliant derivatives, querying datasets programmatically, and preparing data for BIDS-Apps processing pipelines
+- **[NeuroKit2](../skills/neurokit2/)** - Reproducible physiological time-series research with NeuroKit2 0.2.13 on Python 3.10+: method-aware preprocessing, event/interval analysis, multimodal alignment, variability, and complexity across ECG, PPG, EDA, respiration, EMG, EOG, and selected EEG workflows. It is a research and educational toolbox, not for diagnosis, treatment, patient-monitoring alarms, or medical-device validation/certification
+- **[Neuropixels-Analysis](../skills/neuropixels-analysis/)** - Comprehensive toolkit for analyzing Neuropixels high-density neural recordings using SpikeInterface, Allen Institute, and International Brain Laboratory (IBL) best practices. Supports the full workflow from raw data to publication-ready curated units. Key features include: data loading from SpikeGLX, Open Ephys, and NWB formats, preprocessing pipelines (highpass filtering, phase shift correction for Neuropixels 1.0, bad channel detection, common average referencing), motion/drift estimation and correction (nonrigid_fast_and_accurate, nonrigid_accurate, and DREDge presets), spike sorting integration (Kilosort4 GPU, SpykingCircus2, Tridesclous2, Mountainsort5 CPU), comprehensive postprocessing (waveform extraction, template computation, spike amplitudes, correlograms, unit locations), quality metrics computation (SNR, ISI violations, presence ratio, amplitude cutoff, drift metrics), automated curation using Allen Institute and IBL criteria with configurable thresholds, model-based curation with pretrained UnitRefine classifiers (noise/neural and SUA/MUA) loaded from Hugging Face via spikeinterface.curation, AI-assisted visual curation for uncertain units using vision-language models, and export to Phy for manual review or NWB for sharing. Supports Neuropixels 1.0 (960 electrodes, 384 channels) and Neuropixels 2.0 (single and 4-shank configurations). Use cases: extracellular electrophysiology analysis, spike sorting from silicon probes, neural population recordings, systems neuroscience research, unit quality assessment, publication-ready neural data processing, and integration of AI-assisted curation for borderline units
 
 ### Protein Engineering & Design
-- **Adaptyv** - Cloud laboratory platform for automated protein testing and validation. Submit protein sequences via API or web interface and receive experimental results in approximately 21 days. Supports multiple assay types including binding assays (biolayer interferometry for protein-target interactions, KD/kon/koff measurements), expression testing (quantify protein expression levels in E. coli, mammalian, yeast, or insect cells), thermostability measurements (DSF and CD for Tm determination and thermal stability profiling), and enzyme activity assays (kinetic parameters, substrate specificity, inhibitor testing). Includes computational optimization tools for pre-screening sequences: NetSolP/SoluProt for solubility prediction, SolubleMPNN for sequence redesign to improve expression, ESM for sequence likelihood scoring, ipTM (AlphaFold-Multimer) for interface stability assessment, and pSAE for aggregation risk quantification. Platform features automated workflows from expression through purification to assay execution with quality control, webhook notifications for experiment completion, batch submission support for high-throughput screening, and comprehensive results with kinetic parameters, confidence metrics, and raw data access. Use cases: antibody affinity maturation, therapeutic protein developability assessment, enzyme engineering and optimization, protein stability improvement, AI-driven protein design validation, library screening for expression and function, lead optimization with experimental feedback, and integration of computational design with wet-lab validation in iterative design-build-test-learn cycles
-- **ESM (Evolutionary Scale Modeling)** - Protein language models from EvolutionaryScale/Biohub for protein design, structure prediction, and representation learning. Covers current `esm` SDK workflows for local ESM3/ESMC open models, Forge-hosted ESM3 and ESMC inference with `ESM_API_KEY`, Biohub-hosted ESMC embeddings, and ESMFold2 all-atom structure prediction. Use cases: novel protein design, sequence/structure co-design, protein embeddings, function annotation, variant generation, and directed evolution workflows
-- **Glycoengineering** - Analyze and engineer protein glycosylation. Scan sequences for N-glycosylation sequons (N-X-S/T), predict O-glycosylation hotspots, and access curated glycoengineering tools (NetOGlyc, GlycoShield, GlycoWorkbench). Use for glycoprotein engineering, therapeutic antibody optimization, and vaccine design
-- **Molecular Dynamics** - Run and analyze molecular dynamics simulations with OpenMM and MDAnalysis. Set up protein/small molecule systems, define force fields, run energy minimization and production MD, analyze trajectories (RMSD, RMSF, contact maps, free energy surfaces). Use for structural biology, drug binding studies, and biophysics research
-- **Tamarind** - Run a large catalog of open-source molecular design and structural biology tools on the Tamarind Bio managed-GPU cloud via its REST API (`x-api-key` header) or MCP server — no local GPUs required. Covers structure prediction (AlphaFold, Boltz-2, Chai-1, ESMFold2), protein/binder/de novo design (RFdiffusion, ProteinMPNN/LigandMPNN, BoltzGen, BindCraft), antibody and nanobody design, humanization and developability, protein-ligand docking (DiffDock, AutoDock Vina) and binding-affinity prediction, MSA generation, and molecular dynamics — all through one uniform job API with single and batch submission and tool chaining (design → fold → score). Discovers tools and schemas live (`GET /tools`, MCP `getAvailableTools`/`getJobSchema`) and reads the key from `TAMARIND_API_KEY`. There is no official Python SDK — use plain `requests` or the MCP server. Use cases: cloud structure prediction and protein design without provisioning hardware, high-throughput batch characterization of sequences/designs, and chaining design-fold-score pipelines
+- **[Adaptyv](../skills/adaptyv/)** - Cloud laboratory platform for automated protein testing and validation. Submit protein sequences via API or web interface and receive experimental results in approximately 21 days. Supports multiple assay types including binding assays (biolayer interferometry for protein-target interactions, KD/kon/koff measurements), expression testing (quantify protein expression levels in E. coli, mammalian, yeast, or insect cells), thermostability measurements (DSF and CD for Tm determination and thermal stability profiling), and enzyme activity assays (kinetic parameters, substrate specificity, inhibitor testing). Includes computational optimization tools for pre-screening sequences: NetSolP/SoluProt for solubility prediction, SolubleMPNN for sequence redesign to improve expression, ESM for sequence likelihood scoring, ipTM (AlphaFold-Multimer) for interface stability assessment, and pSAE for aggregation risk quantification. Platform features automated workflows from expression through purification to assay execution with quality control, webhook notifications for experiment completion, batch submission support for high-throughput screening, and comprehensive results with kinetic parameters, confidence metrics, and raw data access. Use cases: antibody affinity maturation, therapeutic protein developability assessment, enzyme engineering and optimization, protein stability improvement, AI-driven protein design validation, library screening for expression and function, lead optimization with experimental feedback, and integration of computational design with wet-lab validation in iterative design-build-test-learn cycles
+- **[ESM (Evolutionary Scale Modeling)](../skills/esm/)** - Protein language models from EvolutionaryScale/Biohub for protein design, structure prediction, and representation learning. Covers current `esm` SDK workflows for local ESM3/ESMC open models, Forge-hosted ESM3 and ESMC inference with `ESM_API_KEY`, Biohub-hosted ESMC embeddings, and ESMFold2 all-atom structure prediction. Use cases: novel protein design, sequence/structure co-design, protein embeddings, function annotation, variant generation, and directed evolution workflows
+- **[Glycoengineering](../skills/glycoengineering/)** - Analyze and engineer protein glycosylation. Scan sequences for N-glycosylation sequons (N-X-S/T), predict O-glycosylation hotspots, and access curated glycoengineering tools (NetOGlyc, GlycoShield, GlycoWorkbench). Use for glycoprotein engineering, therapeutic antibody optimization, and vaccine design
+- **[Molecular Dynamics](../skills/molecular-dynamics/)** - Run and analyze molecular dynamics simulations with OpenMM and MDAnalysis. Set up protein/small molecule systems, define force fields, run energy minimization and production MD, analyze trajectories (RMSD, RMSF, contact maps, free energy surfaces). Use for structural biology, drug binding studies, and biophysics research
+- **[Tamarind](../skills/tamarind/)** - Run a large catalog of open-source molecular design and structural biology tools on the Tamarind Bio managed-GPU cloud via its REST API (`x-api-key` header) or MCP server — no local GPUs required. Covers structure prediction (AlphaFold, Boltz-2, Chai-1, ESMFold2), protein/binder/de novo design (RFdiffusion, ProteinMPNN/LigandMPNN, BoltzGen, BindCraft), antibody and nanobody design, humanization and developability, protein-ligand docking (DiffDock, AutoDock Vina) and binding-affinity prediction, MSA generation, and molecular dynamics — all through one uniform job API with single and batch submission and tool chaining (design → fold → score). Discovers tools and schemas live (`GET /tools`, MCP `getAvailableTools`/`getJobSchema`) and reads the key from `TAMARIND_API_KEY`. There is no official Python SDK — use plain `requests` or the MCP server. Use cases: cloud structure prediction and protein design without provisioning hardware, high-throughput batch characterization of sequences/designs, and chaining design-fold-score pipelines
 
 ### Machine Learning & Deep Learning
-- **aeon** - Comprehensive scikit-learn compatible Python toolkit for time series machine learning providing state-of-the-art algorithms across 7 domains: classification (13 algorithm categories including ROCKET variants, deep learning with InceptionTime/ResNet/FCN, distance-based with DTW/ERP/LCSS, shapelet-based, dictionary methods like BOSS/WEASEL, and hybrid ensembles HIVECOTE), regression (9 categories mirroring classification approaches), clustering (k-means/k-medoids with temporal distances, deep learning autoencoders, spectral methods), forecasting (ARIMA, ETS, Theta, Threshold Autoregressive, TCN, DeepAR), anomaly detection (STOMP/MERLIN matrix profile, clustering-based CBLOF/KMeans, isolation methods, copula-based COPOD), segmentation (ClaSP, FLUSS, HMM, binary segmentation), and similarity search (MASS algorithm, STOMP motif discovery, approximate nearest neighbors). Includes 40+ distance metrics (elastic: DTW/DDTW/WDTW/Shape-DTW, edit-based: ERP/EDR/LCSS/TWE/MSM, lock-step: Euclidean/Manhattan), extensive transformations (ROCKET/MiniRocket/MultiRocket for features, Catch22/TSFresh for statistics, SAX/PAA for symbolic representation, shapelet transforms, wavelets, matrix profile), 20+ deep learning architectures (FCN, ResNet, InceptionTime, TCN, autoencoders with attention mechanisms), comprehensive benchmarking tools (UCR/UEA archives with 100+ datasets, published results repository, statistical testing), and performance-optimized implementations using numba. Features progressive model complexity from fast baselines (MiniRocket: <1 second training, 0.95+ accuracy on many benchmarks) to state-of-the-art ensembles (HIVECOTE V2), GPU acceleration support, and extensive visualization utilities. Use cases: physiological signal classification (ECG, EEG), industrial sensor monitoring, financial forecasting, change point detection, pattern discovery, activity recognition from wearables, predictive maintenance, climate time series analysis, and any sequential data requiring specialized temporal modeling beyond standard ML
-- **Cirq** - Google quantum computing framework for designing, simulating, and running quantum circuits. Best suited for targeting Google Quantum AI hardware, designing noise-aware circuits, and running quantum characterization experiments. Provides low-level circuit design, noise modeling, and hardware-specific optimizations. For IBM hardware use Qiskit; for quantum ML with autodiff use PennyLane; for physics simulations use QuTiP
-- **PufferLib** - High-performance reinforcement learning library achieving 1M-4M steps/second through optimized vectorization, native multi-agent support, and efficient PPO training (PuffeRL). Use this skill for RL training on any environment (Gymnasium, PettingZoo, Atari, Procgen), creating custom PufferEnv environments, developing policies (CNN, LSTM, multi-input architectures), optimizing parallel simulation performance, or scaling multi-agent systems. Includes Ocean suite (20+ environments), seamless framework integration with automatic space flattening, zero-copy vectorization with shared memory buffers, distributed training support, and comprehensive reference guides for training workflows, environment development, vectorization optimization, policy architectures, and third-party integrations
-- **PyMC** - Comprehensive Python library for Bayesian statistical modeling and probabilistic programming. Provides intuitive syntax for building probabilistic models, advanced MCMC sampling algorithms (NUTS, nutpie, Metropolis-Hastings, Slice sampling), variational inference methods (ADVI, SVGD), Gaussian processes, time series distributions, and model comparison tools (WAIC, LOO). Features include automatic differentiation via PyTensor, compiled backends such as Numba/JAX, parallel sampling, model diagnostics and convergence checking, PyMC 6 data-container prediction workflows, and integration with ArviZ for visualization and analysis. Supports hierarchical models, mixture models, survival analysis, and custom distributions. Use cases: Bayesian data analysis, uncertainty quantification, A/B testing, time series forecasting, hierarchical modeling, and probabilistic machine learning
-- **PyMOO** - Python framework for multi-objective optimization using evolutionary algorithms. Provides implementations of state-of-the-art algorithms including NSGA-II, NSGA-III, MOEA/D, SPEA2, and reference-point based methods. Features include: support for constrained and unconstrained optimization, multiple problem types (continuous, discrete, mixed-variable), performance indicators (hypervolume, IGD, GD), visualization tools (Pareto front plots, convergence plots), and parallel evaluation support. Supports custom problem definitions, algorithm configuration, and result analysis. Designed for engineering design, parameter optimization, and any problem requiring optimization of multiple conflicting objectives simultaneously. Use cases: multi-objective optimization problems, Pareto-optimal solution finding, engineering design optimization, and research in evolutionary computation
-- **PyTorch Lightning** - Deep learning framework (lightning 2.6+) that organizes PyTorch code to eliminate boilerplate while maintaining full flexibility. Automates training workflows (epoch/batch iteration, optimizer steps, gradient management, checkpointing), supports multi-GPU/TPU training with DDP/FSDP/DeepSpeed strategies, includes LightningModule for model organization, Trainer for automation, LightningDataModule for data pipelines, callbacks for extensibility, and integrations with TensorBoard, Wandb, and MLflow for experiment tracking
-- **PennyLane** - Cross-platform Python library for quantum computing, quantum machine learning, and quantum chemistry. Enables building and training quantum circuits with automatic differentiation, seamless integration with PyTorch/JAX/NumPy, and device-independent execution across simulators and quantum hardware (IBM, Amazon Braket, Google, Rigetti, IonQ). Key features include: quantum circuit construction with QNodes (quantum functions with automatic differentiation), 100+ quantum gates and operations (Pauli, Hadamard, rotation, controlled gates), circuit templates and layers for common ansatze (StronglyEntanglingLayers, BasicEntanglerLayers, UCCSD for chemistry), gradient computation methods (parameter-shift rule for hardware, backpropagation for simulators, adjoint differentiation), quantum chemistry module (molecular Hamiltonian construction, VQE for ground state energy, differentiable Hartree-Fock solver), ML framework integration (TorchLayer for PyTorch models, JAX transformations, TensorFlow deprecated), built-in optimizers (Adam, GradientDescent, QNG, Rotosolve), measurement types (expectation values, probabilities, samples, state vectors), device ecosystem (default.qubit simulator, lightning.qubit for performance, hardware plugins for IBM/Braket/Cirq/Rigetti/IonQ), and Catalyst for just-in-time compilation with adaptive circuits. Supports variational quantum algorithms (VQE, QAOA), quantum neural networks, hybrid quantum-classical models, data encoding strategies (angle, amplitude, IQP embeddings), and pulse-level programming. Use cases: variational quantum eigensolver for molecular simulations, quantum circuit machine learning with gradient-based optimization, hybrid quantum-classical neural networks, quantum chemistry calculations with differentiable workflows, quantum algorithm prototyping with hardware-agnostic code, quantum machine learning research with automatic differentiation, and deploying quantum circuits across multiple quantum computing platforms
-- **Qiskit** - Current Qiskit 2.5 workflow for constructing circuits and observables, running V2 Sampler and Estimator PUBs, compiling ISA circuits with BackendV2 targets, simulating locally with statevector primitives or Aer, and executing on IBM Quantum through Runtime 0.48 job, batch, or session modes. Covers secure IBM Quantum Platform authentication, observable layout mapping, parameter sweeps, noise-management and mitigation options, QPY serialization, migration from removed Qiskit 0.x/1.x APIs, Qiskit Algorithms/Nature/Machine Learning/Optimization, modular Qiskit addons, reproducibility, testing, and read-only environment/backend inspection scripts
-- **QuTiP** - Quantum Toolbox in Python for simulating and analyzing quantum mechanical systems. Provides comprehensive tools for both closed (unitary) and open (dissipative) quantum systems including quantum states (kets, bras, density matrices, Fock states, coherent states), quantum operators (creation/annihilation operators, Pauli matrices, angular momentum operators, quantum gates), time evolution solvers (Schrödinger equation with sesolve, Lindblad master equation with mesolve, quantum trajectories with Monte Carlo mcsolve, Bloch-Redfield brmesolve, Floquet methods for periodic Hamiltonians), analysis tools (expectation values, entropy measures, fidelity, concurrence, correlation functions, steady state calculations), visualization (Bloch sphere with animations, Wigner functions, Q-functions, Fock distributions, matrix histograms), and advanced methods (Hierarchical Equations of Motion for non-Markovian dynamics, permutational invariance for identical particles, stochastic solvers, superoperators). Supports tensor products for composite systems, partial traces, time-dependent Hamiltonians, multiple dissipation channels, and parallel processing. Includes extensive documentation, tutorials, and examples. Use cases: quantum optics simulations (cavity QED, photon statistics), quantum computing (gate operations, circuit dynamics), open quantum systems (decoherence, dissipation), quantum information theory (entanglement dynamics, quantum channels), condensed matter physics (spin chains, many-body systems), and general quantum mechanics research and education
-- **scikit-learn** - Industry-standard Python library for classical machine learning (current stable **1.8.0**, Python 3.11–3.14, free-threaded CPython wheels). Provides comprehensive supervised learning (classification: Logistic Regression, SVM, Decision Trees, Random Forests with 17+ variants, Gradient Boosting with XGBoost-compatible HistGradientBoosting, Naive Bayes, KNN, Neural Networks/MLP; regression: Linear, Ridge, Lasso, ElasticNet, SVR, ensemble methods), unsupervised learning (clustering: K-Means, DBSCAN, native HDBSCAN since 1.3, OPTICS, Agglomerative/Hierarchical, Spectral, Gaussian Mixture Models, BIRCH, MeanShift; dimensionality reduction: PCA, Kernel PCA, t-SNE, Isomap, LLE, MDS, ClassicalMDS, NMF, TruncatedSVD, FastICA, LDA; outlier detection: IsolationForest, LocalOutlierFactor, OneClassSVM), data preprocessing (scaling: StandardScaler, MinMaxScaler, RobustScaler; encoding: OneHotEncoder, OrdinalEncoder, LabelEncoder; imputation: SimpleImputer, KNNImputer, IterativeImputer; feature engineering: PolynomialFeatures, KBinsDiscretizer, text vectorization with CountVectorizer/TfidfVectorizer), model evaluation (cross-validation: KFold, StratifiedKFold, TimeSeriesSplit, GroupKFold; hyperparameter tuning: GridSearchCV, RandomizedSearchCV, HalvingGridSearchCV; metrics: 30+ evaluation metrics for classification/regression/clustering including accuracy, precision, recall, F1, ROC-AUC, MSE, RMSE via root_mean_squared_error, R², silhouette score), and Pipeline/ColumnTransformer for production-ready workflows. Features consistent API (fit/predict/transform), extensive documentation, integration with NumPy/pandas/SciPy, joblib persistence, and scikit-learn-compatible ecosystem (XGBoost, LightGBM, CatBoost, imbalanced-learn, UMAP via umap-learn). Optimized implementations using Cython/OpenMP for performance. Use cases: predictive modeling, customer segmentation, anomaly detection, feature engineering, model selection/validation, text classification, image classification (with feature extraction), time series forecasting (with preprocessing), medical diagnosis, fraud detection, recommendation systems, and any tabular data ML task requiring interpretable models or established algorithms
-- **scikit-survival** - Survival analysis and time-to-event modeling with censored data. Built on scikit-learn, provides Cox proportional hazards models (CoxPHSurvivalAnalysis, CoxnetSurvivalAnalysis with elastic net regularization), ensemble methods (Random Survival Forests, Gradient Boosting), Survival Support Vector Machines (linear and kernel), non-parametric estimators (Kaplan-Meier, Nelson-Aalen), competing risks analysis, and specialized evaluation metrics (concordance index, time-dependent AUC, Brier score). Handles right-censored data, integrates with scikit-learn pipelines, and supports feature selection and hyperparameter tuning via cross-validation
-- **SHAP** - Model explanation and auditing with SHAP 0.52 (Python 3.12+). Covers the modern `shap.Explanation` API, output-aware Tree/Linear/Permutation/Partition/Deep explainers, background-data and masker selection, multiclass slicing, additivity validation, cohort analysis, text/image workflows, and current bar/beeswarm/waterfall/scatter visualizations. Emphasizes reproducible reports and clear limits: predictive attributions are not causal effects, fairness certificates, or recourse
-- **Stable Baselines3** - PyTorch-based reinforcement learning library (current stable **2.8.0**, Python 3.10+, PyTorch >= 2.3). Provides reliable implementations of RL algorithms (PPO, SAC, DQN, TD3, DDPG, A2C, HER, RecurrentPPO via SB3-Contrib). Use this skill for training RL agents on standard or custom Gymnasium environments, implementing callbacks for monitoring and control, using vectorized environments for parallel training, creating custom environments with proper Gymnasium API implementation, and integrating with deep RL workflows. Includes comprehensive training templates, evaluation utilities, algorithm selection guidance (on-policy vs off-policy, continuous vs discrete actions), support for multi-input policies (dict observations), goal-conditioned learning with HER, and integration with TensorBoard for experiment tracking
-- **statsmodels** - Statistical modeling and econometrics (OLS, GLM, logit/probit, ARIMA, time series forecasting, hypothesis testing, diagnostics)
-- **TimesFM Forecasting** - Zero-shot time series forecasting with Google's TimesFM foundation model. Use for any univariate time series (sales, sensors, energy, vitals, weather) without training a custom model. Supports CSV/DataFrame/array inputs with point forecasts and prediction intervals. Includes a preflight system checker script to verify RAM/GPU before first use
-- **Torch Geometric** - PyTorch Geometric (PyG 2.7.x) for graph neural networks on PyTorch 2.6+. Covers `Data`/`HeteroData`, 60+ conv layers (GCN, GAT, GraphSAGE, GIN), node/link/graph classification, heterogeneous graphs, neighbor sampling (`NeighborLoader`, `LinkNeighborLoader`), OGB and built-in datasets, custom dataset loading, GNN explainability, and scaling via DDP, Lightning, and `torch.compile`
-- **Transformers** - Hugging Face Transformers 5.x (PyTorch) for NLP, computer vision, audio, and multimodal tasks. Provides 1M+ pre-trained Hub models via pipelines (text-classification, NER, QA, summarization, translation, text-generation, image-classification, object-detection, ASR, VQA), Trainer API fine-tuning with mixed precision and distributed training, flexible text generation (greedy, beam search, sampling), and Auto classes for architecture selection (BERT, GPT, T5, ViT, BART, etc.)
-- **UMAP-learn** - Python implementation of Uniform Manifold Approximation and Projection (UMAP) for dimensionality reduction and manifold learning. Provides fast, scalable nonlinear dimensionality reduction that preserves both local and global structure of high-dimensional data. Key features include: support for both supervised and unsupervised dimensionality reduction, ability to handle mixed data types, integration with scikit-learn API, and efficient implementation using numba for performance. Produces low-dimensional embeddings (typically 2D or 3D) suitable for visualization and downstream analysis. Often outperforms t-SNE in preserving global structure while maintaining local neighborhoods. Use cases: data visualization, feature extraction, preprocessing for machine learning, single-cell data analysis, and exploratory data analysis of high-dimensional datasets
+- **[aeon](../skills/aeon/)** - Comprehensive scikit-learn compatible Python toolkit for time series machine learning providing state-of-the-art algorithms across 7 domains: classification (13 algorithm categories including ROCKET variants, deep learning with InceptionTime/ResNet/FCN, distance-based with DTW/ERP/LCSS, shapelet-based, dictionary methods like BOSS/WEASEL, and hybrid ensembles HIVECOTE), regression (9 categories mirroring classification approaches), clustering (k-means/k-medoids with temporal distances, deep learning autoencoders, spectral methods), forecasting (ARIMA, ETS, Theta, Threshold Autoregressive, TCN, DeepAR), anomaly detection (STOMP/MERLIN matrix profile, clustering-based CBLOF/KMeans, isolation methods, copula-based COPOD), segmentation (ClaSP, FLUSS, HMM, binary segmentation), and similarity search (MASS algorithm, STOMP motif discovery, approximate nearest neighbors). Includes 40+ distance metrics (elastic: DTW/DDTW/WDTW/Shape-DTW, edit-based: ERP/EDR/LCSS/TWE/MSM, lock-step: Euclidean/Manhattan), extensive transformations (ROCKET/MiniRocket/MultiRocket for features, Catch22/TSFresh for statistics, SAX/PAA for symbolic representation, shapelet transforms, wavelets, matrix profile), 20+ deep learning architectures (FCN, ResNet, InceptionTime, TCN, autoencoders with attention mechanisms), comprehensive benchmarking tools (UCR/UEA archives with 100+ datasets, published results repository, statistical testing), and performance-optimized implementations using numba. Features progressive model complexity from fast baselines (MiniRocket: <1 second training, 0.95+ accuracy on many benchmarks) to state-of-the-art ensembles (HIVECOTE V2), GPU acceleration support, and extensive visualization utilities. Use cases: physiological signal classification (ECG, EEG), industrial sensor monitoring, financial forecasting, change point detection, pattern discovery, activity recognition from wearables, predictive maintenance, climate time series analysis, and any sequential data requiring specialized temporal modeling beyond standard ML
+- **[Cirq](../skills/cirq/)** - Google quantum computing framework for designing, simulating, and running quantum circuits. Best suited for targeting Google Quantum AI hardware, designing noise-aware circuits, and running quantum characterization experiments. Provides low-level circuit design, noise modeling, and hardware-specific optimizations. For IBM hardware use Qiskit; for quantum ML with autodiff use PennyLane; for physics simulations use QuTiP
+- **[PufferLib](../skills/pufferlib/)** - Version-aware guidance for reinforcement-learning environments, vectorization, policies, PuffeRL training, evaluation, and safe checkpoint review. Published `pufferlib==3.0.0` provides the Python/Gymnasium/PettingZoo emulation, vector, and Torch PuffeRL surface; the redesigned native 4.0 source line is incompatible and removes those 3.0 modules. Start with CPU-only synthetic tools and audit native builds, plug-ins, assets, checkpoints, GPU use, logging, and network effects before execution
+- **[PyMC](../skills/pymc/)** - Comprehensive Python library for Bayesian statistical modeling and probabilistic programming. Provides intuitive syntax for building probabilistic models, advanced MCMC sampling algorithms (NUTS, nutpie, Metropolis-Hastings, Slice sampling), variational inference methods (ADVI, SVGD), Gaussian processes, time series distributions, and model comparison tools (WAIC, LOO). Features include automatic differentiation via PyTensor, compiled backends such as Numba/JAX, parallel sampling, model diagnostics and convergence checking, PyMC 6 data-container prediction workflows, and integration with ArviZ for visualization and analysis. Supports hierarchical models, mixture models, survival analysis, and custom distributions. Use cases: Bayesian data analysis, uncertainty quantification, A/B testing, time series forecasting, hierarchical modeling, and probabilistic machine learning
+- **[PyMOO](../skills/pymoo/)** - Python framework for multi-objective optimization using evolutionary algorithms. Provides implementations of state-of-the-art algorithms including NSGA-II, NSGA-III, MOEA/D, SPEA2, and reference-point based methods. Features include: support for constrained and unconstrained optimization, multiple problem types (continuous, discrete, mixed-variable), performance indicators (hypervolume, IGD, GD), visualization tools (Pareto front plots, convergence plots), and parallel evaluation support. Supports custom problem definitions, algorithm configuration, and result analysis. Designed for engineering design, parameter optimization, and any problem requiring optimization of multiple conflicting objectives simultaneously. Use cases: multi-objective optimization problems, Pareto-optimal solution finding, engineering design optimization, and research in evolutionary computation
+- **[PyTorch Lightning](../skills/pytorch-lightning/)** - Deep learning framework (lightning 2.6+) that organizes PyTorch code to eliminate boilerplate while maintaining full flexibility. Automates training workflows (epoch/batch iteration, optimizer steps, gradient management, checkpointing), supports multi-GPU/TPU training with DDP/FSDP/DeepSpeed strategies, includes LightningModule for model organization, Trainer for automation, LightningDataModule for data pipelines, callbacks for extensibility, and integrations with TensorBoard, Wandb, and MLflow for experiment tracking
+- **[PennyLane](../skills/pennylane/)** - Cross-platform Python library for quantum computing, quantum machine learning, and quantum chemistry. Enables building and training quantum circuits with automatic differentiation, seamless integration with PyTorch/JAX/NumPy, and device-independent execution across simulators and quantum hardware (IBM, Amazon Braket, Google, Rigetti, IonQ). Key features include: quantum circuit construction with QNodes (quantum functions with automatic differentiation), 100+ quantum gates and operations (Pauli, Hadamard, rotation, controlled gates), circuit templates and layers for common ansatze (StronglyEntanglingLayers, BasicEntanglerLayers, UCCSD for chemistry), gradient computation methods (parameter-shift rule for hardware, backpropagation for simulators, adjoint differentiation), quantum chemistry module (molecular Hamiltonian construction, VQE for ground state energy, differentiable Hartree-Fock solver), ML framework integration (TorchLayer for PyTorch models, JAX transformations, TensorFlow deprecated), built-in optimizers (Adam, GradientDescent, QNG, Rotosolve), measurement types (expectation values, probabilities, samples, state vectors), device ecosystem (default.qubit simulator, lightning.qubit for performance, hardware plugins for IBM/Braket/Cirq/Rigetti/IonQ), and Catalyst for just-in-time compilation with adaptive circuits. Supports variational quantum algorithms (VQE, QAOA), quantum neural networks, hybrid quantum-classical models, data encoding strategies (angle, amplitude, IQP embeddings), and pulse-level programming. Use cases: variational quantum eigensolver for molecular simulations, quantum circuit machine learning with gradient-based optimization, hybrid quantum-classical neural networks, quantum chemistry calculations with differentiable workflows, quantum algorithm prototyping with hardware-agnostic code, quantum machine learning research with automatic differentiation, and deploying quantum circuits across multiple quantum computing platforms
+- **[Qiskit](../skills/qiskit/)** - Current Qiskit 2.5 workflow for constructing circuits and observables, running V2 Sampler and Estimator PUBs, compiling ISA circuits with BackendV2 targets, simulating locally with statevector primitives or Aer, and executing on IBM Quantum through Runtime 0.48 job, batch, or session modes. Covers secure IBM Quantum Platform authentication, observable layout mapping, parameter sweeps, noise-management and mitigation options, QPY serialization, migration from removed Qiskit 0.x/1.x APIs, Qiskit Algorithms/Nature/Machine Learning/Optimization, modular Qiskit addons, reproducibility, testing, and read-only environment/backend inspection scripts
+- **[QuTiP](../skills/qutip/)** - Simulate and audit closed and open quantum-system models with QuTiP 5.3.0 on Python 3.11+, including deterministic evolution, trajectories, steady states, spectra, phase-space methods, Floquet, HEOM, and permutational-invariance workflows. Physical assumptions, tensor dimensions, truncation, tolerances, and numerical convergence must be explicit. QuTiP is a local simulation toolbox, not a hardware-execution SDK; circuit/control features live in separate QuTiP family packages
+- **[scikit-learn](../skills/scikit-learn/)** - Industry-standard Python library for classical machine learning (current stable **1.8.0**, Python 3.11–3.14, free-threaded CPython wheels). Provides comprehensive supervised learning (classification: Logistic Regression, SVM, Decision Trees, Random Forests with 17+ variants, Gradient Boosting with XGBoost-compatible HistGradientBoosting, Naive Bayes, KNN, Neural Networks/MLP; regression: Linear, Ridge, Lasso, ElasticNet, SVR, ensemble methods), unsupervised learning (clustering: K-Means, DBSCAN, native HDBSCAN since 1.3, OPTICS, Agglomerative/Hierarchical, Spectral, Gaussian Mixture Models, BIRCH, MeanShift; dimensionality reduction: PCA, Kernel PCA, t-SNE, Isomap, LLE, MDS, ClassicalMDS, NMF, TruncatedSVD, FastICA, LDA; outlier detection: IsolationForest, LocalOutlierFactor, OneClassSVM), data preprocessing (scaling: StandardScaler, MinMaxScaler, RobustScaler; encoding: OneHotEncoder, OrdinalEncoder, LabelEncoder; imputation: SimpleImputer, KNNImputer, IterativeImputer; feature engineering: PolynomialFeatures, KBinsDiscretizer, text vectorization with CountVectorizer/TfidfVectorizer), model evaluation (cross-validation: KFold, StratifiedKFold, TimeSeriesSplit, GroupKFold; hyperparameter tuning: GridSearchCV, RandomizedSearchCV, HalvingGridSearchCV; metrics: 30+ evaluation metrics for classification/regression/clustering including accuracy, precision, recall, F1, ROC-AUC, MSE, RMSE via root_mean_squared_error, R², silhouette score), and Pipeline/ColumnTransformer for production-ready workflows. Features consistent API (fit/predict/transform), extensive documentation, integration with NumPy/pandas/SciPy, joblib persistence, and scikit-learn-compatible ecosystem (XGBoost, LightGBM, CatBoost, imbalanced-learn, UMAP via umap-learn). Optimized implementations using Cython/OpenMP for performance. Use cases: predictive modeling, customer segmentation, anomaly detection, feature engineering, model selection/validation, text classification, image classification (with feature extraction), time series forecasting (with preprocessing), medical diagnosis, fraud detection, recommendation systems, and any tabular data ML task requiring interpretable models or established algorithms
+- **[scikit-survival](../skills/scikit-survival/)** - Build, evaluate, and audit right-censored or competing-risk workflows with scikit-survival 0.28.0 on Python 3.11+, including leakage-safe preprocessing, Cox/Coxnet, trees, forests, boosting, SVMs, probability prediction, censoring-aware metrics, and nonparametric cumulative incidence. Built-in competing-risk support is not Fine-Gray regression; model output is not clinical advice, causal evidence, or proof of clinical utility
+- **[SHAP](../skills/shap/)** - Model explanation and auditing with SHAP 0.52 (Python 3.12+). Covers the modern `shap.Explanation` API, output-aware Tree/Linear/Permutation/Partition/Deep explainers, background-data and masker selection, multiclass slicing, additivity validation, cohort analysis, text/image workflows, and current bar/beeswarm/waterfall/scatter visualizations. Emphasizes reproducible reports and clear limits: predictive attributions are not causal effects, fairness certificates, or recourse
+- **[Stable Baselines3](../skills/stable-baselines3/)** - PyTorch-based reinforcement learning library (current stable **2.8.0**, Python 3.10+, PyTorch >= 2.3). Provides reliable implementations of RL algorithms (PPO, SAC, DQN, TD3, DDPG, A2C, HER, RecurrentPPO via SB3-Contrib). Use this skill for training RL agents on standard or custom Gymnasium environments, implementing callbacks for monitoring and control, using vectorized environments for parallel training, creating custom environments with proper Gymnasium API implementation, and integrating with deep RL workflows. Includes comprehensive training templates, evaluation utilities, algorithm selection guidance (on-policy vs off-policy, continuous vs discrete actions), support for multi-input policies (dict observations), goal-conditioned learning with HER, and integration with TensorBoard for experiment tracking
+- **[statsmodels](../skills/statsmodels/)** - Statistical modeling and econometrics (OLS, GLM, logit/probit, ARIMA, time series forecasting, hypothesis testing, diagnostics)
+- **[TimesFM Forecasting](../skills/timesfm-forecasting/)** - Zero-shot time series forecasting with Google's TimesFM foundation model. Use for any univariate time series (sales, sensors, energy, vitals, weather) without training a custom model. Supports CSV/DataFrame/array inputs with point forecasts and prediction intervals. Includes a preflight system checker script to verify RAM/GPU before first use
+- **[Torch Geometric](../skills/torch-geometric/)** - PyTorch Geometric (PyG 2.7.x) for graph neural networks on PyTorch 2.6+. Covers `Data`/`HeteroData`, 60+ conv layers (GCN, GAT, GraphSAGE, GIN), node/link/graph classification, heterogeneous graphs, neighbor sampling (`NeighborLoader`, `LinkNeighborLoader`), OGB and built-in datasets, custom dataset loading, GNN explainability, and scaling via DDP, Lightning, and `torch.compile`
+- **[Transformers](../skills/transformers/)** - Hugging Face Transformers 5.x (PyTorch) for NLP, computer vision, audio, and multimodal tasks. Provides 1M+ pre-trained Hub models via pipelines (text-classification, NER, QA, summarization, translation, text-generation, image-classification, object-detection, ASR, VQA), Trainer API fine-tuning with mixed precision and distributed training, flexible text generation (greedy, beam search, sampling), and Auto classes for architecture selection (BERT, GPT, T5, ViT, BART, etc.)
+- **[UMAP-learn](../skills/umap-learn/)** - Python implementation of Uniform Manifold Approximation and Projection (UMAP) for dimensionality reduction and manifold learning. Provides fast, scalable nonlinear dimensionality reduction that preserves both local and global structure of high-dimensional data. Key features include: support for both supervised and unsupervised dimensionality reduction, ability to handle mixed data types, integration with scikit-learn API, and efficient implementation using numba for performance. Produces low-dimensional embeddings (typically 2D or 3D) suitable for visualization and downstream analysis. Often outperforms t-SNE in preserving global structure while maintaining local neighborhoods. Use cases: data visualization, feature extraction, preprocessing for machine learning, single-cell data analysis, and exploratory data analysis of high-dimensional datasets
 
 ### Materials Science & Chemistry
-- **Astropy** - Comprehensive Python library for astronomy and astrophysics providing core functionality for astronomical research and data analysis. Includes coordinate system transformations (ICRS, Galactic, FK5, AltAz), physical units and quantities with automatic dimensional consistency, FITS file operations (reading, writing, manipulating headers and data), cosmological calculations (luminosity distance, lookback time, Hubble parameter, Planck/WMAP models), precise time handling across multiple time scales (UTC, TAI, TT, TDB) and formats (JD, MJD, ISO), table operations with unit support (FITS, CSV, HDF5, VOTable), WCS transformations between pixel and world coordinates, astronomical constants, modeling framework, visualization tools, and statistical functions. Use for celestial coordinate transformations, unit conversions, FITS image/table processing, cosmological distance calculations, barycentric time corrections, catalog cross-matching, and astronomical data analysis
-- **COBRApy** - Python package for constraint-based reconstruction and analysis (COBRA) of metabolic networks. Provides tools for building, manipulating, and analyzing genome-scale metabolic models (GEMs). Key features include: flux balance analysis (FBA) for predicting optimal metabolic fluxes, flux variability analysis (FVA), gene knockout simulations, pathway analysis, model validation, and integration with other COBRA Toolbox formats (SBML, JSON). Supports various optimization objectives (biomass production, ATP production, metabolite production), constraint handling (reaction bounds, gene-protein-reaction associations), and model comparison. Includes utilities for model construction, gap filling, and model refinement. Use cases: metabolic engineering, systems biology, biotechnology applications, understanding cellular metabolism, and predicting metabolic phenotypes
-- **Pymatgen** - Python Materials Genomics (pymatgen) library for materials science computation and analysis. Provides comprehensive tools for crystal structure manipulation, phase diagram construction, electronic structure analysis, and materials property calculations. Key features include: structure objects with symmetry analysis, space group determination, structure matching and comparison, phase diagram generation from formation energies, band structure and density of states analysis, defect calculations, surface and interface analysis, and integration with DFT codes (VASP, Quantum ESPRESSO, ABINIT). Supports Materials Project database integration, structure file I/O (CIF, POSCAR, VASP), and high-throughput materials screening workflows. Use cases: materials discovery, crystal structure analysis, phase stability prediction, electronic structure calculations, and computational materials science research
+- **[Astropy](../skills/astropy/)** - Comprehensive Python library for astronomy and astrophysics providing core functionality for astronomical research and data analysis. Includes coordinate system transformations (ICRS, Galactic, FK5, AltAz), physical units and quantities with automatic dimensional consistency, FITS file operations (reading, writing, manipulating headers and data), cosmological calculations (luminosity distance, lookback time, Hubble parameter, Planck/WMAP models), precise time handling across multiple time scales (UTC, TAI, TT, TDB) and formats (JD, MJD, ISO), table operations with unit support (FITS, CSV, HDF5, VOTable), WCS transformations between pixel and world coordinates, astronomical constants, modeling framework, visualization tools, and statistical functions. Use for celestial coordinate transformations, unit conversions, FITS image/table processing, cosmological distance calculations, barycentric time corrections, catalog cross-matching, and astronomical data analysis
+- **[COBRApy](../skills/cobrapy/)** - Python package for constraint-based reconstruction and analysis (COBRA) of metabolic networks. Provides tools for building, manipulating, and analyzing genome-scale metabolic models (GEMs). Key features include: flux balance analysis (FBA) for predicting optimal metabolic fluxes, flux variability analysis (FVA), gene knockout simulations, pathway analysis, model validation, and integration with other COBRA Toolbox formats (SBML, JSON). Supports various optimization objectives (biomass production, ATP production, metabolite production), constraint handling (reaction bounds, gene-protein-reaction associations), and model comparison. Includes utilities for model construction, gap filling, and model refinement. Use cases: metabolic engineering, systems biology, biotechnology applications, understanding cellular metabolism, and predicting metabolic phenotypes
+- **[Pymatgen](../skills/pymatgen/)** - Analyze, validate, convert, and transform structures and computed materials data with the current split stack: `pymatgen==2026.5.4`, `pymatgen-core==2026.7.16`, and `mp-api==0.46.4` on Python 3.11+. Covers provenance-preserving local phase diagrams, symmetry sensitivity, transformations, and electronic-structure I/O; Materials Project queries are explicitly bounded and require user-approved network access plus the named `MP_API_KEY`
 
 ### Engineering & Simulation
-- **MATLAB/Octave** - Numerical computing environment for matrix operations, data analysis, visualization, and scientific computing. MATLAB is commercial software optimized for matrix operations, while GNU Octave is a free open-source alternative with high compatibility. Key features include: matrix operations (creation, manipulation, linear algebra), comprehensive mathematics (eigenvalues, SVD, FFT, ODEs, optimization, statistics), 2D/3D visualization (plot, surf, contour, with extensive customization), data import/export (CSV, Excel, MAT files, images), programming constructs (functions, scripts, control flow, OOP), signal processing (FFT, filtering, convolution), and Python integration (calling Python from MATLAB and vice versa). Supports vectorized operations for performance, anonymous functions, tables for mixed data types, and cell arrays for heterogeneous data. GNU Octave provides compatibility with most MATLAB scripts with minor differences (comments with #, block terminators like endif, compound operators like +=). Scripts can be executed via `matlab -nodisplay -r "run('script.m'); exit;"` or `octave script.m`. Use cases: numerical simulations, signal processing, image processing, control systems, statistical analysis, algorithm prototyping, data visualization, and any scientific computing task requiring matrix operations or numerical methods
-- **FluidSim** - Object-oriented Python framework for high-performance computational fluid dynamics (CFD) simulations using pseudospectral methods with FFT. Provides solvers for periodic-domain equations including 2D/3D incompressible Navier-Stokes equations (with/without stratification), shallow water equations, and Föppl-von Kármán elastic plate equations. Key features include: Pythran/Transonic compilation for performance comparable to Fortran/C++, MPI parallelization for large-scale simulations, hierarchical parameter configuration with type safety, comprehensive output management (physical fields in HDF5, spatial means, energy/enstrophy spectra, spectral energy budgets), custom forcing mechanisms (time-correlated random forcing, proportional forcing, script-defined forcing), flexible initial conditions (noise, vortex, dipole, Taylor-Green, from file, in-script), online and offline visualization, and integration with ParaView/VisIt for 3D visualization. Supports workflow features including simulation restart/continuation, parametric studies with batch execution, cluster submission integration, and adaptive CFL-based time stepping. Use cases: 2D/3D turbulence studies with energy cascade analysis, stratified oceanic and atmospheric flows with buoyancy effects, geophysical flows with rotation (Coriolis effects), vortex dynamics and fundamental fluid mechanics research, high-resolution direct numerical simulation (DNS), parametric studies exploring parameter spaces, validation studies (Taylor-Green vortex), and any periodic-domain fluid dynamics research requiring HPC-grade performance with Python flexibility
-- **SimPy** - Process-based discrete-event simulation framework for modeling systems with processes, queues, and resource contention (manufacturing, service operations, network traffic, logistics). Supports generator-based process definition, multiple resource types (Resource, PriorityResource, PreemptiveResource, Container, Store), event-driven scheduling, process interaction mechanisms (signaling, interruption, parallel/sequential execution), real-time simulation synchronized with wall-clock time, and comprehensive monitoring capabilities for utilization, wait times, and queue statistics
-- **SymPy** - Symbolic mathematics in Python for exact computation using mathematical symbols rather than numerical approximations. Provides comprehensive support for symbolic algebra (simplification, expansion, factorization), calculus (derivatives, integrals, limits, series), equation solving (algebraic, differential, systems of equations), matrices and linear algebra (eigenvalues, decompositions, solving linear systems), physics (classical mechanics with Lagrangian/Hamiltonian formulations, quantum mechanics, vector analysis, units), number theory (primes, factorization, modular arithmetic, Diophantine equations), geometry (2D/3D analytic geometry), combinatorics (permutations, combinations, partitions, group theory), logic and sets, statistics (probability distributions, random variables), special functions (gamma, Bessel, orthogonal polynomials), and code generation (lambdify to NumPy/SciPy functions, C/Fortran code generation, LaTeX output for documentation). Emphasizes exact arithmetic using rational numbers and symbolic representations, supports assumptions for improved simplification (positive, real, integer), integrates seamlessly with NumPy/SciPy through lambdify for fast numerical evaluation, and enables symbolic-to-numeric pipelines for scientific computing workflows
+- **[MATLAB/Octave](../skills/matlab/)** - Build, review, migrate, and safely plan numerical workflows against proprietary MATLAB R2026a or the distinct free GNU Octave 11.3.0 surface. Covers arrays, tables/timetables, tests, projects, graphics, MAT-file inventory, and explicit Python interoperability. Bundled Python 3.11+ tools are local static/dry-run helpers and never launch either runtime; do not assume product, toolbox, license, API, numerical, or graphics equivalence
+- **[FluidSim](../skills/fluidsim/)** - Plan, configure, inspect, restart, and analyze bounded FluidSim 0.9.0 pseudospectral CFD simulations, with the verified FluidFFT 0.4.5/pyFFTW 0.15.1 stack. Requires explicit equations, units, solver parameters, convergence tests, and CPU/RAM/disk/wall-time bounds; MPI/native FFT execution must follow an approved site scheduler/toolchain workflow. A completed or stable run is not proof of numerical convergence or physical validity
+- **[SimPy](../skills/simpy/)** - Build, test, and analyze bounded process-based discrete-event simulations with SimPy 4.1.2 on Python 3.8+, including events, resources, interrupts, monitoring, replications, warm-up, and reproducible output analysis. SimPy supplies scheduling primitives but does not select a valid conceptual model, input distribution, estimand, run length, replication count, or causal interpretation
+- **[SymPy](../skills/sympy/)** - Symbolic mathematics in Python for exact computation using mathematical symbols rather than numerical approximations. Provides comprehensive support for symbolic algebra (simplification, expansion, factorization), calculus (derivatives, integrals, limits, series), equation solving (algebraic, differential, systems of equations), matrices and linear algebra (eigenvalues, decompositions, solving linear systems), physics (classical mechanics with Lagrangian/Hamiltonian formulations, quantum mechanics, vector analysis, units), number theory (primes, factorization, modular arithmetic, Diophantine equations), geometry (2D/3D analytic geometry), combinatorics (permutations, combinations, partitions, group theory), logic and sets, statistics (probability distributions, random variables), special functions (gamma, Bessel, orthogonal polynomials), and code generation (lambdify to NumPy/SciPy functions, C/Fortran code generation, LaTeX output for documentation). Emphasizes exact arithmetic using rational numbers and symbolic representations, supports assumptions for improved simplification (positive, real, integer), integrates seamlessly with NumPy/SciPy through lambdify for fast numerical evaluation, and enables symbolic-to-numeric pipelines for scientific computing workflows
 
 ### Data Analysis & Visualization
-- **Dask** - Parallel and distributed computing for larger-than-RAM pandas/NumPy workflows (dask 2026.x; Python 3.10+). Provides lazy DataFrames (expression-based query planning since 2025.1), Arrays, Bags, delayed tasks, and Futures via dask.distributed. Scales from multi-core laptops to clusters with a diagnostic dashboard. PyArrow 16+ required for DataFrame I/O; cloud paths need s3fs/gcsfs. Use for parallel file processing, out-of-core ETL, and scaling existing pandas/NumPy code. For single-machine out-of-core analytics without a cluster, consider vaex; for in-memory speed use polars
-- **GeoMaster** - Comprehensive geospatial science skill covering remote sensing, GIS, spatial analysis, machine learning for earth observation, and 30+ scientific domains. Supports satellite imagery processing (Sentinel, Landsat, MODIS, SAR, hyperspectral), vector and raster data operations, spatial statistics, point cloud processing, network analysis, cloud-native workflows (STAC, COG, Planetary Computer), and 8 programming languages (Python, R, Julia, JavaScript, C++, Java, Go, Rust) with 500+ code examples. Use for remote sensing workflows, GIS analysis, spatial ML, Earth observation data processing, terrain analysis, hydrological modeling, marine spatial analysis, atmospheric science, and any geospatial computation task
-- **GeoPandas** - Python library extending pandas for working with geospatial vector data including shapefiles, GeoJSON, and GeoPackage files. Provides GeoDataFrame and GeoSeries data structures combining geometric data with tabular attributes for spatial analysis. Key features include: reading/writing spatial file formats (Shapefile, GeoJSON, GeoPackage, PostGIS, Parquet) with Arrow acceleration for 2-4x faster I/O, geometric operations (buffer, simplify, centroid, convex hull, affine transformations) through Shapely integration, spatial analysis (spatial joins with predicates like intersects/contains/within, nearest neighbor joins, overlay operations for union/intersection/difference, dissolve for aggregation, clipping), coordinate reference system (CRS) management (setting CRS, reprojecting between coordinate systems, UTM estimation), and visualization (static choropleth maps with matplotlib, interactive maps with folium, multi-layer mapping, classification schemes with mapclassify). Supports spatial indexing for performance, filtering during read operations (bbox, mask, SQL WHERE), and integration with cartopy for cartographic projections. Use cases: spatial data manipulation, buffer analysis, spatial joins between datasets, dissolving boundaries, calculating areas/distances in projected CRS, reprojecting coordinate systems, creating choropleth maps, converting between spatial file formats, PostGIS database integration, and geospatial data analysis workflows
-- **Matplotlib** - Comprehensive Python plotting library for creating publication-quality static, animated, and interactive visualizations. Provides extensive customization options for creating figures, subplots, axes, and annotations. Key features include: support for multiple plot types (line, scatter, bar, histogram, contour, 3D, and many more), extensive customization (colors, fonts, styles, layouts), multiple backends (PNG, PDF, SVG, interactive backends), LaTeX integration for mathematical notation, and integration with NumPy and pandas. Includes specialized modules (pyplot for MATLAB-like interface, artist layer for fine-grained control, backend layer for rendering). Supports complex multi-panel figures, color maps, legends, and annotations. Use cases: scientific figure creation, data visualization, exploratory data analysis, publication graphics, and any application requiring high-quality plots
-- **NetworkX** - Comprehensive toolkit for creating, analyzing, and visualizing complex networks and graphs. Supports four graph types (Graph, DiGraph, MultiGraph, MultiDiGraph) with nodes as any hashable objects and rich edge attributes. Provides 100+ algorithms including shortest paths (Dijkstra, Bellman-Ford, A*), centrality measures (degree, betweenness, closeness, eigenvector, PageRank), clustering (coefficients, triangles, transitivity), community detection (modularity-based, label propagation, Girvan-Newman), connectivity analysis (components, cuts, flows), tree algorithms (MST, spanning trees), matching, graph coloring, isomorphism, and traversal (DFS, BFS). Includes 50+ graph generators for classic (complete, cycle, wheel), random (Erdős-Rényi, Barabási-Albert, Watts-Strogatz, stochastic block model), lattice (grid, hexagonal, hypercube), and specialized networks. Supports I/O across formats (edge lists, GraphML, GML, JSON, Pajek, GEXF, DOT) with Pandas/NumPy/SciPy integration. Visualization capabilities include 8+ layout algorithms (spring/force-directed, circular, spectral, Kamada-Kawai), customizable node/edge appearance, interactive visualizations with Plotly/PyVis, and publication-quality figure generation. Use cases: social network analysis, biological networks (protein-protein interactions, gene regulatory networks, metabolic pathways), transportation systems, citation networks, knowledge graphs, web structure analysis, infrastructure networks, and any domain involving pairwise relationships requiring structural analysis or graph-based modeling
-- **Polars** - High-performance DataFrame library written in Rust with Python bindings for fast data manipulation, ETL, analytics, and pandas migration. Provides expression-based transformations, lazy query optimization, automatic parallel execution, streaming out-of-core processing, Arrow interoperability, and optional GPU execution. Supports common data sources and formats including CSV, Parquet, JSON/NDJSON, Excel, Arrow IPC, cloud object storage, databases, and BigQuery. Use cases: large-scale data processing, memory-conscious analytical pipelines, feature engineering, and high-performance DataFrame workflows
-- **Seaborn** - Statistical data visualization with dataset-oriented interface, automatic confidence intervals, publication-quality themes, colorblind-safe palettes, and comprehensive support for exploratory analysis, distribution comparisons, correlation matrices, regression plots, and multi-panel figures
-- **Vaex** - High-performance Python library for lazy, out-of-core DataFrames to process and visualize tabular datasets larger than available RAM. Processes over a billion rows per second through memory-mapped files (HDF5, Apache Arrow), lazy evaluation, and virtual columns (zero memory overhead). Provides instant file opening, efficient aggregations across billions of rows, interactive visualizations without sampling, machine learning pipelines with transformers (scalers, encoders, PCA), and seamless integration with pandas/NumPy/Arrow. Includes comprehensive ML framework (vaex.ml) with feature scaling, categorical encoding, dimensionality reduction, and integration with scikit-learn/XGBoost/LightGBM/CatBoost. Supports distributed computing via Dask, asynchronous operations, and state management for production deployment. Use cases: processing gigabyte to terabyte datasets, fast statistical aggregations on massive data, visualizing billion-row datasets, ML pipelines on big data, converting between data formats, and working with astronomical, financial, or scientific large-scale datasets
+- **[Dask](../skills/dask/)** - Parallel and distributed computing for larger-than-RAM pandas/NumPy workflows (dask 2026.x; Python 3.10+). Provides lazy DataFrames (expression-based query planning since 2025.1), Arrays, Bags, delayed tasks, and Futures via dask.distributed. Scales from multi-core laptops to clusters with a diagnostic dashboard. PyArrow 16+ required for DataFrame I/O; cloud paths need s3fs/gcsfs. Use for parallel file processing, out-of-core ETL, and scaling existing pandas/NumPy code. For single-machine out-of-core analytics without a cluster, consider vaex; for in-memory speed use polars
+- **[GeoMaster](../skills/geomaster/)** - Comprehensive geospatial science skill covering remote sensing, GIS, spatial analysis, machine learning for earth observation, and 30+ scientific domains. Supports satellite imagery processing (Sentinel, Landsat, MODIS, SAR, hyperspectral), vector and raster data operations, spatial statistics, point cloud processing, network analysis, cloud-native workflows (STAC, COG, Planetary Computer), and 8 programming languages (Python, R, Julia, JavaScript, C++, Java, Go, Rust) with 500+ code examples. Use for remote sensing workflows, GIS analysis, spatial ML, Earth observation data processing, terrain analysis, hydrological modeling, marine spatial analysis, atmospheric science, and any geospatial computation task
+- **[GeoPandas](../skills/geopandas/)** - Guidance and local audits for planar vector workflows using stable GeoPandas 1.1.4 on Python 3.10+, not unreleased 1.2 documentation. Covers GeoSeries/GeoDataFrame construction, CRS inspection and transformation, spatial joins/overlays, geometry validity, distance/area caveats, Arrow-aware I/O, and bounded metadata-first review of local vector files; avoid treating longitude/latitude as planar distance or trusting unreviewed remote/database inputs
+- **[Matplotlib](../skills/matplotlib/)** - Comprehensive Python plotting library for creating publication-quality static, animated, and interactive visualizations. Provides extensive customization options for creating figures, subplots, axes, and annotations. Key features include: support for multiple plot types (line, scatter, bar, histogram, contour, 3D, and many more), extensive customization (colors, fonts, styles, layouts), multiple backends (PNG, PDF, SVG, interactive backends), LaTeX integration for mathematical notation, and integration with NumPy and pandas. Includes specialized modules (pyplot for MATLAB-like interface, artist layer for fine-grained control, backend layer for rendering). Supports complex multi-panel figures, color maps, legends, and annotations. Use cases: scientific figure creation, data visualization, exploratory data analysis, publication graphics, and any application requiring high-quality plots
+- **[NetworkX](../skills/networkx/)** - Comprehensive toolkit for creating, analyzing, and visualizing complex networks and graphs. Supports four graph types (Graph, DiGraph, MultiGraph, MultiDiGraph) with nodes as any hashable objects and rich edge attributes. Provides 100+ algorithms including shortest paths (Dijkstra, Bellman-Ford, A*), centrality measures (degree, betweenness, closeness, eigenvector, PageRank), clustering (coefficients, triangles, transitivity), community detection (modularity-based, label propagation, Girvan-Newman), connectivity analysis (components, cuts, flows), tree algorithms (MST, spanning trees), matching, graph coloring, isomorphism, and traversal (DFS, BFS). Includes 50+ graph generators for classic (complete, cycle, wheel), random (Erdős-Rényi, Barabási-Albert, Watts-Strogatz, stochastic block model), lattice (grid, hexagonal, hypercube), and specialized networks. Supports I/O across formats (edge lists, GraphML, GML, JSON, Pajek, GEXF, DOT) with Pandas/NumPy/SciPy integration. Visualization capabilities include 8+ layout algorithms (spring/force-directed, circular, spectral, Kamada-Kawai), customizable node/edge appearance, interactive visualizations with Plotly/PyVis, and publication-quality figure generation. Use cases: social network analysis, biological networks (protein-protein interactions, gene regulatory networks, metabolic pathways), transportation systems, citation networks, knowledge graphs, web structure analysis, infrastructure networks, and any domain involving pairwise relationships requiring structural analysis or graph-based modeling
+- **[Polars](../skills/polars/)** - High-performance DataFrame library written in Rust with Python bindings for fast data manipulation, ETL, analytics, and pandas migration. Provides expression-based transformations, lazy query optimization, automatic parallel execution, streaming out-of-core processing, Arrow interoperability, and optional GPU execution. Supports common data sources and formats including CSV, Parquet, JSON/NDJSON, Excel, Arrow IPC, cloud object storage, databases, and BigQuery. Use cases: large-scale data processing, memory-conscious analytical pipelines, feature engineering, and high-performance DataFrame workflows
+- **[Seaborn](../skills/seaborn/)** - Statistical data visualization with dataset-oriented interface, automatic confidence intervals, publication-quality themes, colorblind-safe palettes, and comprehensive support for exploratory analysis, distribution comparisons, correlation matrices, regression plots, and multi-panel figures
+- **[Vaex](../skills/vaex/)** - High-performance Python library for lazy, out-of-core DataFrames to process and visualize tabular datasets larger than available RAM. Processes over a billion rows per second through memory-mapped files (HDF5, Apache Arrow), lazy evaluation, and virtual columns (zero memory overhead). Provides instant file opening, efficient aggregations across billions of rows, interactive visualizations without sampling, machine learning pipelines with transformers (scalers, encoders, PCA), and seamless integration with pandas/NumPy/Arrow. Includes comprehensive ML framework (vaex.ml) with feature scaling, categorical encoding, dimensionality reduction, and integration with scikit-learn/XGBoost/LightGBM/CatBoost. Supports distributed computing via Dask, asynchronous operations, and state management for production deployment. Use cases: processing gigabyte to terabyte datasets, fast statistical aggregations on massive data, visualizing billion-row datasets, ML pipelines on big data, converting between data formats, and working with astronomical, financial, or scientific large-scale datasets
 
 ### Phylogenetics & Evolutionary Biology
-- **ETE Toolkit** - Analyze, manipulate, compare, annotate, and visualize existing phylogenetic or hierarchical trees with ETE 4. Covers explicit Newick/Nexus parsing and Newick serialization, topology edits, rooting and `TreePattern` matching, Robinson-Foulds comparison and distance matrices, `PhyloTree` duplication/speciation inference and gene-tree reconciliation, local NCBI and GTDB taxonomy databases, browser-based SmartView exploration, and PNG/PDF/SVG rendering. Includes ETE 3→4 migration guidance and validated command-line helpers. Use dedicated alignment and tree-inference tools before ETE when starting from raw sequences
-- **Phylogenetics** - Build and analyze phylogenetic trees using MAFFT (multiple alignment), IQ-TREE 2 (maximum likelihood), and FastTree (fast NJ/ML). Visualize with ETE 4 or FigTree. Use for evolutionary analysis, microbial genomics, viral phylodynamics, protein family analysis, and molecular clock studies
+- **[ETE Toolkit](../skills/etetoolkit/)** - Analyze, manipulate, compare, annotate, and visualize existing phylogenetic or hierarchical trees with ETE 4. Covers explicit Newick/Nexus parsing and Newick serialization, topology edits, rooting and `TreePattern` matching, Robinson-Foulds comparison and distance matrices, `PhyloTree` duplication/speciation inference and gene-tree reconciliation, local NCBI and GTDB taxonomy databases, browser-based SmartView exploration, and PNG/PDF/SVG rendering. Includes ETE 3→4 migration guidance and validated command-line helpers. Use dedicated alignment and tree-inference tools before ETE when starting from raw sequences
+- **[Phylogenetics](../skills/phylogenetics/)** - Build and analyze phylogenetic trees using MAFFT (multiple alignment), IQ-TREE 2 (maximum likelihood), and FastTree (fast NJ/ML). Visualize with ETE 4 or FigTree. Use for evolutionary analysis, microbial genomics, viral phylodynamics, protein family analysis, and molecular clock studies
 
-### Multi-omics & AI Agent Frameworks
-- **Pi Agent** - Build with and use Pi, the minimal terminal coding harness. Covers installing and configuring Pi, authenticating providers, adding custom models and providers, creating Pi skills/extensions/packages/themes/prompt templates, embedding Pi through the Node/TypeScript SDK, integrating over RPC or JSON event streams, parsing session JSONL files, and building custom TUI components. Use cases: building custom agent UIs, exposing Pi from another process or language, wiring private model gateways, packaging reusable Pi workflows, and helping agents operate Pi as a platform rather than only a CLI
-- **HypoGeniC** - Automated hypothesis generation and testing using large language models to accelerate scientific discovery. Provides three frameworks: HypoGeniC (data-driven hypothesis generation from observational data), HypoRefine (synergistic approach combining literature insights with empirical patterns through an agentic system), and Union methods (mechanistic combination of literature and data-driven hypotheses). Features iterative refinement that improves hypotheses by learning from challenging examples, Redis caching for API cost reduction, and customizable YAML-based prompt templates. Includes command-line tools for generation (hypogenic_generation) and testing (hypogenic_inference). Research applications have demonstrated 14.19% accuracy improvement in AI-content detection and 7.44% in deception detection. Use cases: deception detection in reviews, AI-generated content identification, mental stress detection, exploratory research without existing literature, hypothesis-driven analysis in novel domains, and systematic exploration of competing explanations
+### Agent Frameworks
+- **[Pi Agent](../skills/pi-agent/)** - Build with and use Pi, the minimal terminal coding harness. Covers installing and configuring Pi, authenticating providers, adding custom models and providers, creating Pi skills/extensions/packages/themes/prompt templates, embedding Pi through the Node/TypeScript SDK, integrating over RPC or JSON event streams, parsing session JSONL files, and building custom TUI components. Use cases: building custom agent UIs, exposing Pi from another process or language, wiring private model gateways, packaging reusable Pi workflows, and helping agents operate Pi as a platform rather than only a CLI
 
 ### Autonomous Research & Optimization Frameworks
-- **Arbor** - Autonomous Optimization (AO) via Hypothesis Tree Refinement (HTR), adapted from the Arbor framework (Jin et al., 2026; RUC-NLPIR/Arbor). Iteratively improves a real artifact — a codebase, model-training recipe, agent/search harness, data-synthesis pipeline, or prompt — against a measurable objective over many experiment cycles, without overfitting to the feedback signal. The agent acts as a long-lived **coordinator** that maintains a persistent **hypothesis tree** (each node binds a falsifiable hypothesis, a distilled reusable insight, and a pointer to the artifact version that realizes it) and runs the six-step loop Observe → Ideate → Select → Dispatch → Backpropagate → Decide; short-lived **executor** subagents each test one hypothesis in an isolated git worktree and return a dev score, factual result, distilled insight, and branch reference. A **held-out merge gate** promotes a candidate to the best artifact only when its improvement transfers to a test evaluator the search never optimized against, separating exploratory dev gains from verified progress. Insights are propagated up the tree so local findings become direction-level constraints and global priors, and falsified branches are pruned with their reasons recorded as negative constraints. Ships a `tree.py` state manager (init/observe/add-node/set-evidence/propagate/prune/merge/status/validate) that keeps the tree consistent and auditable, plus reference docs on the HTR methodology, the executor brief, the final-report template, and how to run the upstream `arbor` CLI instead. Use cases: "improve my model's eval score in fewer steps", optimizing an agent or search harness for higher pass rate/accuracy, tuning a data-generation pipeline judged by downstream model behavior, MLE-bench / Kaggle-style "improve the submission" tasks, and any long-horizon "make this artifact better and don't just memorize the dev set" workflow that needs structured, branching exploration with an audit trail
+- **[Arbor](../skills/arbor/)** - Autonomous Optimization (AO) via Hypothesis Tree Refinement (HTR), adapted from the Arbor framework (Jin et al., 2026; RUC-NLPIR/Arbor). Iteratively improves a real artifact — a codebase, model-training recipe, agent/search harness, data-synthesis pipeline, or prompt — against a measurable objective over many experiment cycles, without overfitting to the feedback signal. The agent acts as a long-lived **coordinator** that maintains a persistent **hypothesis tree** (each node binds a falsifiable hypothesis, a distilled reusable insight, and a pointer to the artifact version that realizes it) and runs the six-step loop Observe → Ideate → Select → Dispatch → Backpropagate → Decide; short-lived **executor** subagents each test one hypothesis in an isolated git worktree and return a dev score, factual result, distilled insight, and branch reference. A **held-out merge gate** promotes a candidate to the best artifact only when its improvement transfers to a test evaluator the search never optimized against, separating exploratory dev gains from verified progress. Insights are propagated up the tree so local findings become direction-level constraints and global priors, and falsified branches are pruned with their reasons recorded as negative constraints. Ships a `tree.py` state manager (init/observe/add-node/set-evidence/propagate/prune/merge/status/validate) that keeps the tree consistent and auditable, plus reference docs on the HTR methodology, the executor brief, the final-report template, and how to run the upstream `arbor` CLI instead. Use cases: "improve my model's eval score in fewer steps", optimizing an agent or search harness for higher pass rate/accuracy, tuning a data-generation pipeline judged by downstream model behavior, MLE-bench / Kaggle-style "improve the submission" tasks, and any long-horizon "make this artifact better and don't just memorize the dev set" workflow that needs structured, branching exploration with an audit trail
 
 ### Scientific Communication & Publishing
-- **BGPT Paper Search** - Search scientific papers and retrieve structured experimental data extracted from full-text studies via the BGPT MCP server. Returns 25+ fields per paper including methods, results, sample sizes, quality scores, and conclusions. Use for literature reviews, evidence synthesis, and finding experimental details not available in abstracts alone
-- **pyzotero** - Python client for the Zotero Web API v3. Programmatically manage Zotero reference libraries: retrieve, create, update, and delete items, collections, tags, and attachments. Export citations as BibTeX, CSL-JSON, and formatted bibliography HTML. Supports user and group libraries, local mode for offline access, paginated retrieval with `everything()`, full-text content indexing, saved search management, and file upload/download. Optional CLI and built-in MCP server (pyzotero 1.12+) for searching local Zotero 7 libraries including full-text PDF search and Semantic Scholar integration. Use cases: building research automation pipelines that integrate with Zotero, bulk importing references, exporting bibliographies programmatically, managing large reference collections, syncing library metadata, enriching bibliographic data, and connecting LLM agents to a local Zotero library.
-- **Citation Management** - Comprehensive citation management for academic research. Search Google Scholar and PubMed for papers, extract accurate metadata from multiple sources (CrossRef, PubMed, arXiv), validate citations, and generate properly formatted BibTeX entries. Features include converting DOIs, PMIDs, or arXiv IDs to BibTeX, cleaning and formatting bibliography files, finding highly cited papers, checking for duplicates, and ensuring consistent citation formatting. Use cases: building bibliographies for manuscripts, verifying citation accuracy, citation deduplication, and maintaining reference databases
-- **Generate Image** - AI-powered image generation and editing for scientific illustrations, schematics, and visualizations using OpenRouter's image generation models. Supports multiple models including google/gemini-3.6-flash (high quality, recommended default) and black-forest-labs/flux.2-pro (fast, high quality). Key features include: text-to-image generation from detailed prompts, image editing capabilities (modify existing images with natural language instructions), automatic base64 encoding/decoding, PNG output with configurable paths, and comprehensive error handling. Requires OpenRouter API key (via .env file or environment variable). Use cases: generating scientific diagrams and illustrations, creating publication-quality figures, editing existing images (changing colors, adding elements, removing backgrounds), producing schematics for papers and presentations, visualizing experimental setups, creating graphical abstracts, and generating conceptual illustrations for scientific communication
-- **Infographics** - Create professional infographics using Nano Banana Pro AI with smart iterative refinement. Uses Gemini 3.6 Flash for quality review. Integrates research-lookup and web search for accurate data. Supports 10 infographic types, 8 industry styles, and colorblind-safe palettes
-- **LaTeX Posters** - Create professional research posters in LaTeX using beamerposter, tikzposter, or baposter. Support for conference presentations, academic posters, and scientific communication with layout design, color schemes, multi-column formats, figure integration, and poster-specific best practices. Features compliance with conference size requirements (A0, A1, 36×48"), complex multi-column layouts, and integration of figures, tables, equations, and citations. Use cases: conference poster sessions, thesis defenses, symposia presentations, and research group templates
-- **Market Research Reports** - Generate comprehensive market research reports (50+ pages) in the style of top consulting firms (McKinsey, BCG, Gartner). Features professional LaTeX formatting, extensive visual generation, deep integration with research-lookup for data gathering, and multi-framework strategic analysis including Porter's Five Forces, PESTLE, SWOT, TAM/SAM/SOM, and BCG Matrix. Use cases: investment decisions, strategic planning, competitive landscape analysis, market sizing, and market entry evaluation
-- **PPTX Posters** - Create professional research posters using PowerPoint/HTML formats for researchers who prefer WYSIWYG tools over LaTeX. Features design principles, layout templates, quality checklists, and export guidance for poster sessions. Use cases: conference posters when LaTeX is not preferred, quick poster creation, and collaborative poster design
-- **Scientific Schematics** - Create publication-quality scientific diagrams using Nano Banana Pro AI with smart iterative refinement. Uses Gemini 3.6 Flash for quality review with document-type-specific thresholds (journal: 8.5/10, conference: 8.0/10, poster: 7.0/10). Specializes in neural network architectures, system diagrams, flowcharts, biological pathways, and complex scientific visualizations. Features natural language input, automatic quality assessment, and publication-ready output. Use cases: creating figures for papers, generating workflow diagrams, visualizing experimental designs, and producing graphical abstracts
-- **Scientific Slides** - Build slide decks and presentations for research talks using PowerPoint and LaTeX Beamer. Features slide structure, design templates, timing guidance, and visual validation. Emphasizes visual engagement with minimal text, research-backed content with proper citations, and story-driven narrative. Use cases: conference presentations, academic seminars, thesis defenses, grant pitches, and professional talks
-- **Venue Templates** - Access comprehensive LaTeX templates, formatting requirements, and submission guidelines for major scientific publication venues (Nature, Science, PLOS, IEEE, ACM), academic conferences (NeurIPS, ICML, CVPR, CHI), research posters, and grant proposals (NSF, NIH, DOE, DARPA). Provides ready-to-use templates and detailed specifications for successful academic submissions. Use cases: manuscript preparation, conference papers, research posters, and grant proposals with venue-specific formatting
+- **[BGPT Paper Search](../skills/bgpt-paper-search/)** - Search scientific papers and retrieve structured experimental data extracted from full-text studies via the BGPT MCP server. Returns 25+ fields per paper including methods, results, sample sizes, quality scores, and conclusions. Use for literature reviews, evidence synthesis, and finding experimental details not available in abstracts alone
+- **[pyzotero](../skills/pyzotero/)** - Python client for the Zotero Web API v3. Programmatically manage Zotero reference libraries: retrieve, create, update, and delete items, collections, tags, and attachments. Export citations as BibTeX, CSL-JSON, and formatted bibliography HTML. Supports user and group libraries, local mode for offline access, paginated retrieval with `everything()`, full-text content indexing, saved search management, and file upload/download. Optional CLI and built-in MCP server (pyzotero 1.12+) for searching local Zotero 7 libraries including full-text PDF search and Semantic Scholar integration. Use cases: building research automation pipelines that integrate with Zotero, bulk importing references, exporting bibliographies programmatically, managing large reference collections, syncing library metadata, enriching bibliographic data, and connecting LLM agents to a local Zotero library.
+- **[Citation Management](../skills/citation-management/)** - Comprehensive citation management for academic research. Search Google Scholar and PubMed for papers, extract accurate metadata from multiple sources (CrossRef, PubMed, arXiv), validate citations, and generate properly formatted BibTeX entries. Features include converting DOIs, PMIDs, or arXiv IDs to BibTeX, cleaning and formatting bibliography files, finding highly cited papers, checking for duplicates, and ensuring consistent citation formatting. Use cases: building bibliographies for manuscripts, verifying citation accuracy, citation deduplication, and maintaining reference databases
+- **[Generate Image](../skills/generate-image/)** - AI-powered image generation and editing for scientific illustrations, schematics, and visualizations using OpenRouter's image generation models. Supports multiple models including google/gemini-3.6-flash (high quality, recommended default) and black-forest-labs/flux.2-pro (fast, high quality). Key features include: text-to-image generation from detailed prompts, image editing capabilities (modify existing images with natural language instructions), automatic base64 encoding/decoding, PNG output with configurable paths, and comprehensive error handling. Requires OpenRouter API key (via .env file or environment variable). Use cases: generating scientific diagrams and illustrations, creating publication-quality figures, editing existing images (changing colors, adding elements, removing backgrounds), producing schematics for papers and presentations, visualizing experimental setups, creating graphical abstracts, and generating conceptual illustrations for scientific communication
+- **[Infographics](../skills/infographics/)** - Create professional infographics using Nano Banana Pro AI with smart iterative refinement. Uses Gemini 3.6 Flash for quality review. Integrates research-lookup and web search for accurate data. Supports 10 infographic types, 8 industry styles, and colorblind-safe palettes
+- **[LaTeX Posters](../skills/latex-posters/)** - Create professional research posters in LaTeX using beamerposter, tikzposter, or baposter. Support for conference presentations, academic posters, and scientific communication with layout design, color schemes, multi-column formats, figure integration, and poster-specific best practices. Features compliance with conference size requirements (A0, A1, 36×48"), complex multi-column layouts, and integration of figures, tables, equations, and citations. Use cases: conference poster sessions, thesis defenses, symposia presentations, and research group templates
+- **[Market Research Reports](../skills/market-research-reports/)** - Build evidence-traceable market reports and assumption-driven sizing or forecast scenarios with claim/source IDs, explicit market definitions, TAM/SAM/SOM reconciliation, sensitivity analysis, and auditable report scaffolds. Bundled Python 3.11+ standard-library tools are offline and make no model/image calls; online research needs approval and source-term review. Never imitate an analyst/consulting brand, invent evidence, present one forecast as certain, or provide investment, legal, antitrust, tax, accounting, or regulatory advice
+- **[PPTX Posters](../skills/pptx-posters/)** - Version 2.0 creates and audits a real, editable, one-slide, macro-free `.pptx` from strict author-approved local content and asset manifests using python-pptx 1.0.2. It does not use HTML conversion, network services, external templates, or generated claims. Exact physical size, printer rules, provenance, asset hashes/licenses, accessibility, package security, and approval hashes must pass before generation; final PowerPoint/PDF/printer/author review remains manual
+- **[Scientific Schematics](../skills/scientific-schematics/)** - Create publication-quality scientific diagrams using Nano Banana Pro AI with smart iterative refinement. Uses Gemini 3.6 Flash for quality review with document-type-specific thresholds (journal: 8.5/10, conference: 8.0/10, poster: 7.0/10). Specializes in neural network architectures, system diagrams, flowcharts, biological pathways, and complex scientific visualizations. Features natural language input, automatic quality assessment, and publication-ready output. Use cases: creating figures for papers, generating workflow diagrams, visualizing experimental designs, and producing graphical abstracts
+- **[Scientific Slides](../skills/scientific-slides/)** - Build slide decks and presentations for research talks using PowerPoint and LaTeX Beamer. Features slide structure, design templates, timing guidance, and visual validation. Emphasizes visual engagement with minimal text, research-backed content with proper citations, and story-driven narrative. Use cases: conference presentations, academic seminars, thesis defenses, grant pitches, and professional talks
+- **[Venue Templates](../skills/venue-templates/)** - Access comprehensive LaTeX templates, formatting requirements, and submission guidelines for major scientific publication venues (Nature, Science, PLOS, IEEE, ACM), academic conferences (NeurIPS, ICML, CVPR, CHI), research posters, and grant proposals (NSF, NIH, DOE, DARPA). Provides ready-to-use templates and detailed specifications for successful academic submissions. Use cases: manuscript preparation, conference papers, research posters, and grant proposals with venue-specific formatting
 
 ### Document Processing & Conversion
-- **DOCX** - Create, read, edit, and manipulate Word documents (.docx files). Supports professional documents with formatting like tables of contents, headings, page numbers, letterheads, find-and-replace, tracked changes, comments, and image insertion. Use for reports, memos, letters, templates, and any Word document deliverable
-- **MarkItDown** - Convert trusted local documents, controlled streams, and selected URIs to structure-preserving Markdown with Microsoft MarkItDown 0.1.6. Covers pinned modular installs; PDF, DOCX, PPTX, XLSX/XLS, HTML, CSV, JSON/XML text, EPUB, ZIP, images, audio, YouTube, and notebooks; the canonical `result.markdown` API; safe local/batch/literature workflows; the official vision OCR plugin; Azure Document Intelligence and Content Understanding; custom plugins; and the local MarkItDown MCP server. Includes explicit SSRF, archive, prompt-injection, plugin, credential, external-transcription, and cloud-data controls, plus deterministic local-input helper scripts with an external-service gate
-- **LiteParse** - Fast, local-first document parser (Rust core, Python/Node bindings) for layout-aware text extraction with bounding boxes—no cloud API required. Parses PDFs natively; converts Office and image formats via LibreOffice/ImageMagick when installed. Outputs layout-preserved plain text or structured JSON with per-page `text_items` (coordinates, font metadata, OCR confidence). Built-in Tesseract OCR with optional HTTP OCR servers (EasyOCR/PaddleOCR-compatible API), page subsets, encrypted PDFs, stdin/bytes parsing, and PNG page screenshots for multimodal agents. CLI (`lit parse`, `lit batch-parse`, `lit screenshot`) and Python API (`LiteParse`, `search_items`). Targets liteparse 2.0.0, Python 3.10+. Use when you need spatial grounding for RAG, batch literature ingestion, or agent vision—not for Markdown (MarkItDown) or PDF merge/split/forms (pdf skill)
-- **Markdown & Mermaid Writing** - Comprehensive markdown and Mermaid diagram writing skill for creating scientific documents, reports, analyses, and visualizations. Establishes text-based diagrams as the default documentation standard with full style guides, 24 diagram type references, and 9 document templates
-- **PDF** - Read, create, combine, split, rotate, watermark, encrypt/decrypt, and manipulate PDF files. Supports extracting text/tables, filling forms, OCR on scanned PDFs, image extraction, and format conversion
-- **PPTX** - Create, read, edit, and manipulate PowerPoint presentations (.pptx files). Use for slide decks, pitch decks, extracting text from presentations, editing existing slides, combining or splitting presentations, and working with templates, layouts, speaker notes, and comments
-- **XLSX** - Spreadsheet creation, editing, and analysis with support for formulas, formatting, data analysis, and visualization
+- **[DOCX](../skills/docx/)** - Create, read, edit, and manipulate Word documents (.docx files). Supports professional documents with formatting like tables of contents, headings, page numbers, letterheads, find-and-replace, tracked changes, comments, and image insertion. Use for reports, memos, letters, templates, and any Word document deliverable
+- **[MarkItDown](../skills/markitdown/)** - Convert trusted local documents, controlled streams, and selected URIs to structure-preserving Markdown with Microsoft MarkItDown 0.1.6. Covers pinned modular installs; PDF, DOCX, PPTX, XLSX/XLS, HTML, CSV, JSON/XML text, EPUB, ZIP, images, audio, YouTube, and notebooks; the canonical `result.markdown` API; safe local/batch/literature workflows; the official vision OCR plugin; Azure Document Intelligence and Content Understanding; custom plugins; and the local MarkItDown MCP server. Includes explicit SSRF, archive, prompt-injection, plugin, credential, external-transcription, and cloud-data controls, plus deterministic local-input helper scripts with an external-service gate
+- **[LiteParse](../skills/liteparse/)** - Fast, local-first document parser (Rust core, Python/Node bindings) for layout-aware text extraction with bounding boxes—no cloud API required. Parses PDFs natively; converts Office and image formats via LibreOffice/ImageMagick when installed. Outputs layout-preserved plain text or structured JSON with per-page `text_items` (coordinates, font metadata, OCR confidence). Built-in Tesseract OCR with optional HTTP OCR servers (EasyOCR/PaddleOCR-compatible API), page subsets, encrypted PDFs, stdin/bytes parsing, and PNG page screenshots for multimodal agents. CLI (`lit parse`, `lit batch-parse`, `lit screenshot`) and Python API (`LiteParse`, `search_items`). Targets liteparse 2.0.0, Python 3.10+. Use when you need spatial grounding for RAG, batch literature ingestion, or agent vision—not for Markdown (MarkItDown) or PDF merge/split/forms (pdf skill)
+- **[Markdown & Mermaid Writing](../skills/markdown-mermaid-writing/)** - Comprehensive markdown and Mermaid diagram writing skill for creating scientific documents, reports, analyses, and visualizations. Establishes text-based diagrams as the default documentation standard with full style guides, 24 diagram type references, and 9 document templates
+- **[PDF](../skills/pdf/)** - Read, create, combine, split, rotate, watermark, encrypt/decrypt, and manipulate PDF files. Supports extracting text/tables, filling forms, OCR on scanned PDFs, image extraction, and format conversion
+- **[PPTX](../skills/pptx/)** - Create, read, edit, and manipulate PowerPoint presentations (.pptx files). Use for slide decks, pitch decks, extracting text from presentations, editing existing slides, combining or splitting presentations, and working with templates, layouts, speaker notes, and comments
+- **[XLSX](../skills/xlsx/)** - Spreadsheet creation, editing, and analysis with support for formulas, formatting, data analysis, and visualization
 
 ### Laboratory Automation & Equipment Control
-- **PyLabRobot** - Hardware-agnostic, pure Python SDK for automated and autonomous laboratories. Provides unified interface for controlling liquid handling robots (Hamilton STAR/STARlet, Opentrons OT-2, Tecan EVO), plate readers (BMG CLARIOstar), heater shakers, incubators, centrifuges, pumps, and scales. Key features include: modular resource management system for plates, tips, and containers with hierarchical deck layouts and JSON serialization; comprehensive liquid handling operations (aspirate, dispense, transfer, serial dilutions, plate replication) with automatic tip and volume tracking; backend abstraction enabling hardware-agnostic protocols that work across different robots; ChatterboxBackend for protocol simulation and testing without hardware; browser-based visualizer for real-time 3D deck state visualization; cross-platform support (Windows, macOS, Linux, Raspberry Pi); and integration capabilities for multi-device workflows combining liquid handlers, analytical equipment, and material handling devices. Use cases: automated sample preparation, high-throughput screening, serial dilution protocols, plate reading workflows, laboratory protocol development and validation, robotic liquid handling automation, and reproducible laboratory automation with state tracking and persistence
+- **[PyLabRobot](../skills/pylabrobot/)** - Develop and review PyLabRobot 0.2.1 resources, liquid-handling plans, offline Chatterbox simulations, and supported-device integrations on Python 3.9+. Bundled planning tools are Python 3.11+, local, and make no serial, USB, or network connections. Physical execution requires model-specific extras, configuration, calibration, hazard review, dry runs, explicit user request, and trained-operator approval at the equipment safety gate
 
 ### Tool Discovery & Computational Resources
-- **Autoskill** - Observe the user's screen via the local screenpipe daemon, detect repeated research workflows, match them against existing scientific-agent-skills, and draft new skills (or composition recipes that chain existing ones) for patterns not yet covered. Use when the user asks to analyze their recent work and propose skills based on what they actually do. Requires the screenpipe daemon running locally on port 3030; all detection runs locally and only redacted cluster summaries reach the LLM
-- **Get Available Resources** - Detect available computational resources and generate strategic recommendations for scientific computing tasks at the start of any computationally intensive scientific task. Automatically identifies CPU capabilities, GPU availability (NVIDIA CUDA, AMD ROCm, Apple Silicon Metal), memory constraints, and disk space. Creates JSON file with resource information and recommendations for parallel processing (joblib, multiprocessing), out-of-core computing (Dask, Zarr), GPU acceleration (PyTorch, JAX), or memory-efficient strategies. Use cases: determining optimal computational approaches before data analysis, model training, or large file operations
+- **[Autoskill](../skills/autoskill/)** - Observe the user's screen via the local screenpipe daemon, detect repeated research workflows, match them against existing scientific-agent-skills, and draft new skills (or composition recipes that chain existing ones) for patterns not yet covered. Use when the user asks to analyze their recent work and propose skills based on what they actually do. Requires the screenpipe daemon running locally on port 3030; all detection runs locally and only redacted cluster summaries reach the LLM
+- **[Get Available Resources](../skills/get-available-resources/)** - On explicit request or before a clearly resource-sensitive local workload, detect effective CPU, memory, disk, scheduler, container, and accelerator limits for the current process. Produces a redacted JSON snapshot and conservative planning helpers on Python 3.11+ (optional psutil 7.2.2), without persistent fingerprinting, stress tests, large allocations, environment dumps, device changes, or assuming visible host hardware is usable
 
 ### Research Methodology & Proposal Writing
-- **Paperzilla** - Chat with your agent about Paperzilla projects, recommendations, and canonical papers. Use for recent project recommendations, recommendation triage, canonical paper details, markdown-based summaries, relevance-to-my-research discussions, recommendation feedback, JSON export, and Atom feed URLs
-- **Paper Lookup** - Search 10 academic paper databases via their REST APIs to find research papers, preprints, and scholarly articles. Covers biomedical literature (PubMed, PMC full text), preprint servers (bioRxiv, medRxiv, arXiv), multidisciplinary indexes (OpenAlex, Crossref, Semantic Scholar), open access aggregators (CORE, Unpaywall). Use for searching research papers, finding citations, looking up articles by DOI or PMID, retrieving abstracts or full text, checking open access availability, exploring citation graphs, and systematic literature searches
-- **Research Grants** - Write competitive research proposals for NSF, NIH, DOE, DARPA, and Taiwan NSTC. Features agency-specific formatting, review criteria understanding, budget preparation, broader impacts statements, significance narratives, innovation sections, and compliance with submission requirements (including PAPPG 24-1 and current NIH salary-cap guidance). Covers project descriptions, specific aims, technical narratives, milestone plans, budget justifications, and biosketches. Optional figures via the scientific-schematics skill. Use cases: federal grant applications, resubmissions with reviewer response, multi-institutional collaborations, and preliminary data sections
-- **Research Lookup** - Compile manuscript-ready scholarly evidence with a Parallel-first Search → Extract → optional Research workflow. Academic mode targets 60 verified, deduplicated references by default and produces a bibliography, evidence matrix, claim-to-source map, synthesis of consensus and conflicting evidence, section briefs, coverage diagnostics, and a reproducible search ledger. Parallel Chat remains an explicit, non-default OpenAI-compatible backend, and Perplexity remains an explicit optional fallback. Use cases: manuscript background research, literature verification, methods precedent, discussion context, citation discovery, and research-gap analysis
-- **Scholar Evaluation** - Apply the ScholarEval framework to systematically evaluate scholarly and research work. Provides structured evaluation methodology based on peer-reviewed research assessment criteria for analyzing academic papers, research proposals, literature reviews, and scholarly writing across multiple quality dimensions. Use cases: evaluating research papers for quality and rigor, assessing methodology design, scoring data analysis approaches, benchmarking research quality, and assessing publication readiness
+- **[Paperzilla](../skills/paperzilla/)** - Chat with your agent about Paperzilla projects, recommendations, and canonical papers. Use for recent project recommendations, recommendation triage, canonical paper details, markdown-based summaries, relevance-to-my-research discussions, recommendation feedback, JSON export, and Atom feed URLs
+- **[Paper Lookup](../skills/paper-lookup/)** - Search 10 academic paper databases via their REST APIs to find research papers, preprints, and scholarly articles. Covers biomedical literature (PubMed, PMC full text), preprint servers (bioRxiv, medRxiv, arXiv), multidisciplinary indexes (OpenAlex, Crossref, Semantic Scholar), open access aggregators (CORE, Unpaywall). Use for searching research papers, finding citations, looking up articles by DOI or PMID, retrieving abstracts or full text, checking open access availability, exploring citation graphs, and systematic literature searches
+- **[Research Grants](../skills/research-grants/)** - Write competitive research proposals for NSF, NIH, DOE, DARPA, and Taiwan NSTC. Features agency-specific formatting, review criteria understanding, budget preparation, broader impacts statements, significance narratives, innovation sections, and compliance with submission requirements (including PAPPG 24-1 and current NIH salary-cap guidance). Covers project descriptions, specific aims, technical narratives, milestone plans, budget justifications, and biosketches. Optional figures via the scientific-schematics skill. Use cases: federal grant applications, resubmissions with reviewer response, multi-institutional collaborations, and preliminary data sections
+- **[Research Lookup](../skills/research-lookup/)** - Compile manuscript-ready scholarly evidence with a Parallel-first Search → Extract → optional Research workflow. Academic mode targets 60 verified, deduplicated references by default and produces a bibliography, evidence matrix, claim-to-source map, synthesis of consensus and conflicting evidence, section briefs, coverage diagnostics, and a reproducible search ledger. Parallel Chat remains an explicit, non-default OpenAI-compatible backend, and Perplexity remains an explicit optional fallback. Use cases: manuscript background research, literature verification, methods precedent, discussion context, citation discovery, and research-gap analysis
+- **[Scholar Evaluation](../skills/scholar-evaluation/)** - Version 2.0 provides qualitative-first, evidence-traceable developmental review of scholarly works and audits low-stakes assessment rubrics with optional Python 3.11+ local JSON/CSV quality controls that use no network, credentials, external models, or subprocesses. Optional rubric scores only summarize submitted evidence against predeclared anchors; they are not natural measurements or decision recommendations. Never rank people or use the skill for hiring, promotion, tenure, admissions, funding, prizes, sanctions, or other consequential decisions
 
-### Regulatory & Standards Compliance
-- **ISO 13485 Certification** - Comprehensive toolkit for preparing ISO 13485:2016 certification documentation for medical device Quality Management Systems. Provides gap analysis of existing documentation, templates for all mandatory documents, compliance checklists, and step-by-step documentation creation. Covers 31 required procedures including Quality Manuals, Medical Device Files, and work instructions. Use cases: starting ISO 13485 certification process, conducting gap analysis, creating or updating QMS documentation, preparing for certification audits, transitioning from FDA QSR to QMSR, and harmonizing with EU MDR requirements
+### Regulatory & Standards Evidence Preparation
+- **[ISO 13485 Certification](../skills/iso-13485-certification/)** - Prepare draft ISO 13485 QMS scope, controlled-document scaffolds, traceability, and local evidence manifests, while separating related FDA QMSR, MDSAP, and EU MDR/IVDR evidence boundaries. Bundled Python 3.11+ standard-library checks use bounded local JSON/Markdown only and make no network calls. The skill does not reproduce the standard, perform an audit, determine applicability or compliance, certify a QMS, validate a certificate, or promise an audit result; qualified RA/QA, legal, management, auditor, and certification-body review is required
 
 ## Scientific Thinking & Analysis
 
 ### Analysis & Methodology
-- **Experimental Design** - Design studies before data collection: choosing a design, randomization (simple, block, stratified, cluster), blocking and stratification, factorial and fractional-factorial designs (DOE), screening (Plackett-Burman), response-surface designs (central composite, Box-Behnken), Latin hypercube, crossover/repeated-measures/split-plot/Latin-square, sequential and adaptive designs, and avoiding pseudoreplication. Includes seeded allocation-schedule and DOE-matrix generators (numpy, pandas, pyDOE3). For sample size see Statistical Power; for analysis see Statistical Analysis
-- **Exploratory Data Analysis** - Comprehensive EDA toolkit with automated statistics, visualizations, and insights for any tabular dataset
-- **Hypothesis Generation** - Structured frameworks for generating and evaluating scientific hypotheses
-- **Literature Review** - Systematic literature search and review toolkit with support for multiple scientific databases (PubMed, bioRxiv, Google Scholar), citation management with multiple citation styles (APA, AMA, Vancouver, Chicago, IEEE, Nature, Science), citation verification and deduplication, search strategies (Boolean operators, MeSH terms, field tags), PDF report generation with formatted references, and comprehensive templates for conducting systematic reviews following PRISMA guidelines
-- **Peer Review** - Comprehensive toolkit for conducting high-quality scientific peer review with structured evaluation of methodology, statistics, reproducibility, ethics, and presentation across all scientific disciplines
-- **Scientific Brainstorming** - Conversational brainstorming partner for generating novel research ideas, exploring connections, challenging assumptions, and developing creative approaches through structured ideation workflows
-- **Scientific Critical Thinking** - Tools and approaches for rigorous scientific reasoning and evaluation
-- **Scientific Visualization** - Best practices and templates for creating publication-quality scientific figures with matplotlib and seaborn, including statistical plots with automatic confidence intervals, colorblind-safe palettes, multi-panel figures, heatmaps, and journal-specific formatting
-- **Scientific Writing** - Comprehensive toolkit for writing, structuring, and formatting scientific research papers using IMRAD format, multiple citation styles (APA, AMA, Vancouver, Chicago, IEEE), reporting guidelines (CONSORT, STROBE, PRISMA), effective figures and tables, field-specific terminology, venue-specific structure expectations, and core writing principles for clarity, conciseness, and accuracy across all scientific disciplines
-- **Statistical Analysis** - Guided hypothesis testing, assumption checks, effect sizes, power analysis, and APA reporting (Pingouin, SciPy, statsmodels, PyMC). For dedicated model APIs see statsmodels and pymc skills; for in-depth sample-size/power planning see Statistical Power; for designing the study see Experimental Design
-- **Statistical Power** - Sample-size and statistical power analysis for planning studies: required n, minimum detectable effect (MDE), and power curves for t-tests, ANOVA, proportions, correlation, chi-square, and regression via a unified closed-form interface, plus a Monte Carlo harness for designs with no formula (logistic/Poisson regression, mixed models, cluster-randomized trials, survival). Covers effect-size choice (SESOI), and adjustments for multiplicity, dropout, and clustering (statsmodels, scipy, pingouin, numpy). For laying out the design see Experimental Design; for post-collection analysis see Statistical Analysis
+- **[Experimental Design](../skills/experimental-design/)** - Design studies before data collection: choosing a design, randomization (simple, block, stratified, cluster), blocking and stratification, factorial and fractional-factorial designs (DOE), screening (Plackett-Burman), response-surface designs (central composite, Box-Behnken), Latin hypercube, crossover/repeated-measures/split-plot/Latin-square, sequential and adaptive designs, and avoiding pseudoreplication. Includes seeded allocation-schedule and DOE-matrix generators (numpy, pandas, pyDOE3). For sample size see Statistical Power; for analysis see Statistical Analysis
+- **[Exploratory Data Analysis](../skills/exploratory-data-analysis/)** - Perform bounded, local EDA on explicitly supported authorized files: redacted CSV/TSV/JSON profiles plus optional NumPy, HDF5, FASTA/FASTQ, and basic image metadata inspection. Core CLIs use Python 3.11+ and are network-free; the full optional format stack targets Python 3.12+. Inputs are untrusted: tools never follow embedded instructions, run macros/code/models, print raw records or identifiers, or automatically clean/transform data. Unknown formats fail closed, and bounded samples are never presented as complete validation
+- **[Hypothesis Generation](../skills/hypothesis-generation/)** - Version 2.0 formulates evidence-bounded questions, candidate hypotheses, rival explanations, causal or associational claims, discriminating predictions, measurements, and preregistration-ready plans. All bundled Python 3.11+ standard-library tools are local, deterministic, bounded, non-scoring, and make no network/model/image calls. Hypotheses remain proposals rather than facts; the skill never automatically ranks/selects them, infers causation, supplies clinical advice, bypasses oversight, or fabricates novelty/evidence
+- **[HypoGeniC](../skills/hypogenic/)** - Plan and audit ChicagoHAI `hypogenic==0.3.5`/HypoRefine workflows over labeled text datasets, task configs, hypothesis banks, and HypoBench data. The software proposes candidate textual patterns and task-prediction statistics; held-out accuracy is not experimental confirmation, causal evidence, novelty, or scientific validity. Model providers, Redis, credentials, data, and network use require separate approval, and model calls never start automatically
+- **[Literature Review](../skills/literature-review/)** - Systematic literature search and review toolkit with support for multiple scientific databases (PubMed, bioRxiv, Google Scholar), citation management with multiple citation styles (APA, AMA, Vancouver, Chicago, IEEE, Nature, Science), citation verification and deduplication, search strategies (Boolean operators, MeSH terms, field tags), PDF report generation with formatted references, and comprehensive templates for conducting systematic reviews following PRISMA guidelines
+- **[Peer Review](../skills/peer-review/)** - Version 2.0 prepares evidence-bounded, constructive review drafts for authorized manuscripts, protocols, preprints, or proposals, including reporting-guideline selection, claim/evidence checks, methods/statistics/reproducibility/ethics critique, and response planning. Bundled Python 3.11+ tools process bounded local JSON/CSV/Markdown and make no network, model, image, or external-service calls. Unpublished material is confidential: confirm authorization, venue AI/tool policy, conflicts, competence, and scope; never send content externally or imply an editorial outcome without permission
+- **[Scientific Brainstorming](../skills/scientific-brainstorming/)** - Evidence-aware early-stage ideation using independent generation, structured discussion, explicit assumptions, transparent evaluation, adversarial review, uncertainty, and decision logs. Treat outputs as proposals, not findings; brainstorming cannot validate hypotheses or grant ethics, biosafety, regulatory, or clinical approval. Optional Python 3.11+ tools are local, deterministic, and make no network or model calls
+- **[Scientific Critical Thinking](../skills/scientific-critical-thinking/)** - Tools and approaches for rigorous scientific reasoning and evaluation
+- **[Scientific Visualization](../skills/scientific-visualization/)** - Create and audit truthful, accessible, publication-ready figures with Matplotlib, Seaborn, or Plotly, including multi-panel layouts, uncertainty/missing-data displays, color/contrast review, metadata validation, and export planning. Pinned examples use Python 3.11+; bundled CLIs are network-free, while Plotly/Kaleido v1 static export needs compatible Chrome/Chromium. Preserve raw data and transformations, never hide or invent evidence, verify current journal rules, and do not claim that a palette, DPI, format, or automated report establishes accessibility or compliance
+- **[Scientific Writing](../skills/scientific-writing/)** - Version 2.0 drafts, revises, and audits manuscripts or research reports with explicit evidence provenance, reporting-guideline coverage, authorship accountability, confidentiality controls, and local consistency checks. Every factual/numeric claim maps to human-verified source IDs; AI is not an author and fluency is not evidence. Bundled Python 3.11+ tools are offline, deterministic, and non-submitting; accountable human authors retain scientific decisions and final approval
+- **[Statistical Analysis](../skills/statistical-analysis/)** - Guided hypothesis testing, assumption checks, effect sizes, power analysis, and APA reporting (Pingouin, SciPy, statsmodels, PyMC). For dedicated model APIs see statsmodels and pymc skills; for in-depth sample-size/power planning see Statistical Power; for designing the study see Experimental Design
+- **[Statistical Power](../skills/statistical-power/)** - Sample-size and statistical power analysis for planning studies: required n, minimum detectable effect (MDE), and power curves for t-tests, ANOVA, proportions, correlation, chi-square, and regression via a unified closed-form interface, plus a Monte Carlo harness for designs with no formula (logistic/Poisson regression, mixed models, cluster-randomized trials, survival). Covers effect-size choice (SESOI), and adjustments for multiplicity, dropout, and clustering (statsmodels, scipy, pingouin, numpy). For laying out the design see Experimental Design; for post-collection analysis see Statistical Analysis
 
 ### Decision & Scenario Analysis
-- **Consciousness Council** - Run a multi-perspective Mind Council deliberation on any question, decision, or creative challenge. Use for diverse viewpoints, tough decisions, devil's advocate analysis, exploring problems from multiple angles, or when facing dilemmas, trade-offs, and complex choices with no obvious answer
-- **DHDNA Profiler** - Extract cognitive patterns and thinking fingerprints from any text. Analyze how someone thinks, understand cognitive style, profile writing or speech patterns, compare thinking styles between people, and gain deeper insight into an author's reasoning patterns, decision-making style, or cognitive signature
-- **What-If Oracle** - Run structured What-If scenario analysis with 4–6 branch possibility exploration (best, likely, worst, wild card). Use for speculative what-if questions, strategic forks, contingency planning, and stress-testing decisions before committing
+- **[Consciousness Council](../skills/consciousness-council/)** - Run a multi-perspective Mind Council deliberation on any question, decision, or creative challenge. Use for diverse viewpoints, tough decisions, devil's advocate analysis, exploring problems from multiple angles, or when facing dilemmas, trade-offs, and complex choices with no obvious answer
+- **[DHDNA Profiler](../skills/dhdna-profiler/)** - Extract cognitive patterns and thinking fingerprints from any text. Analyze how someone thinks, understand cognitive style, profile writing or speech patterns, compare thinking styles between people, and gain deeper insight into an author's reasoning patterns, decision-making style, or cognitive signature
+- **[What-If Oracle](../skills/what-if-oracle/)** - Run structured What-If scenario analysis with 4–6 branch possibility exploration (best, likely, worst, wild card). Use for speculative what-if questions, strategic forks, contingency planning, and stress-testing decisions before committing
 
 ### Web Search & Information Retrieval
-- **Exa Search** - Web search and URL content extraction via the Exa API. Use for high-quality web search tuned for scientific and technical content, scholarly filtering via `category="research paper"` plus academic domain allowlists, and batch URL extraction
-- **Parallel Web** - Search the web, extract URL content, and run deep research using the Parallel Chat API and Extract API. Use for web searches, research queries, and general information gathering with synthesized summaries and citations
+- **[Exa Search](../skills/exa-search/)** - Web search and URL content extraction via the Exa API. Use for high-quality web search tuned for scientific and technical content, scholarly filtering via `category="research paper"` plus academic domain allowlists, and batch URL extraction
+- **[Parallel Web](../skills/parallel-web/)** - Search the web, extract URL content, and run deep research using the Parallel Chat API and Extract API. Use for web searches, research queries, and general information gathering with synthesized summaries and citations