NumPy plugins

Execute 58 bioinformatics AI agent skills on genetic files (VCF, FASTQ, h5ad, proteomics) for pharmacogenomics, ancestry inference, scRNA-seq analysis, metagenomics profiling, variant annotation, GWAS, and clinical reporting. Generate reproducible markdown reports, plots, CSV/JSON outputs, and bundled environments via deterministic local Python runs with privacy safeguards.

Automate Chinese research paper workflows: brainstorm topics and structures interactively, synthesize literature from PubMed/arXiv with BibTeX, generate Python data visualizations/stats code using matplotlib/seaborn/scipy, draft evidence-driven chapters, peer-review with checklists, polish bilingual text, and compile publication-ready LaTeX with journal templates.

Analyze survey microdata with weighted pandas DataFrames to compute Gini coefficients, poverty rates, quantiles, and inequality metrics. Impute missing values using ML methods like random forest and XGBoost from donor data. Calibrate weights to population targets with L0 regularization. Enhance datasets like CPS ASEC and run PolicyEngine microsimulations for tax-benefit policy impacts across populations.

Automate end-to-end ML performance investigations: research SOTA papers and architectures, generate phased plans, judge experimental methodologies, profile bottlenecks, run metric-improvement campaigns with atomic git commits, auto-rollback on regressions, and leverage specialist agents for data lifecycle and deep paper analysis.

Automate end-to-end chip and mask layout design in KLayout using AI agents via MCP server: detect van der Waals flakes from microscope images, align GDS to photos, trace contours to polygons, route nanodevice contacts to pads, overlay images, toggle layers, capture visuals, and export GDS from natural language queries.

Generate optimized SQL queries from natural language for BigQuery, PostgreSQL, MySQL, and Snowflake; perform cohort analysis on CSV/Excel user data to compute retention rates, visualize trends, and detect anomalies; evaluate A/B tests with statistical significance, confidence intervals, and launch recommendations.

Bootstrap Claude Code with 17 specialized agents, skills, and hooks to audit/evolve .claude/ configs, engineer/refactor Python code via TDD, profile/optimize ML workloads, generate docs/tests, design systems, diagnose issues, and manage workflows professionally.

Streamline end-to-end data science and ML workflows: frame business problems into ML tasks, preprocess and validate data with quality checks, perform EDA on diverse formats, design and execute experiments with hyperparameter tuning via Optuna and interpretability via SHAP, audit reproducibility and leakage, evaluate model performance and readiness for deployment, generate model cards, and extract structured learnings into docs.

Automate academic research in finance, economics, and real estate: search literature via Corbis, generate and rank novel ideas with heuristics, screen for novelty and journal fit, visualize trends and gaps with Python-generated figures, draft structured reviews and positioning memos, read/summarize papers with a subagent, and audit citations in LaTeX files.

Perform high-performance steady-state analysis on distribution power systems—including power flow, state estimation, and IEC 60909 short-circuit calculations—using Python's power-grid-model library with numpy structured arrays for batch and parallel simulations across 22 component types.

Execute astrophysics analysis workflows: compute JAX-accelerated bandflux for supernova light curves with synthetic data and likelihood templates; process FITS files via photometry, spectroscopy, astrometry, light curve analysis, and cosmological calculations using astropy ecosystem.

Build domain expertise on any topic via zero-context research pipeline: generate question trees, discover and fetch sources to disk, index into .mv2 knowledge store, distill compact artifacts for agent use without raw docs in context. Run Python in isolated Docker sandboxes for safe data analysis, prototyping, and DSPy sub-agents.

Balance class-imbalanced datasets for machine learning workflows using oversampling (SMOTE, ADASYN, Borderline-SMOTE), undersampling, or hybrid techniques applied only to training splits, then save resampled X/y train/test data to an output directory.

Automate end-to-end content creation: collect RSS hotspots and trends, generate viral AI-drafted articles for WeChat Public Account, Xiaohongshu, and Feishu with style imitation and visuals, review for quality/compliance risks, analyze performance from data files with charts, and publish via APIs or Playwright browser automation after one-time credential setup.