100+ bioinformatics Agent Skills: pharmacogenomics, clinical variant interpretation, GWAS and PRS, ancestry, RNA-seq and single-cell, metagenomics, proteomics, methylation clocks, nf-core pipeline wrappers. Many run deterministically from the CLI with demo data and write a reproducibility bundle. Local-first, privacy-focused.
Turn a plain-language scientific problem into a grounded, runnable Claude solution — composed from the cataloged Skills/MCP servers/Plugins, reusing a curated recipe when one fits, and recommending a pre-built autonomous system when that's the right call.
Hypothesis Crucible — agentic cross-corpus hypothesis generation. Mines typed knowledge fragments from literature, structured databases, and raw experimental data; bridges them Swanson-style into novel connections; then runs a falsification gauntlet of independent veto gates that aggressively rejects non-novel, ungrounded, contradicted, or implausible ideas, surfacing only survivors with cited support and a discriminating experimental test.
Medicine Design Software Development: the drug discovery pipeline and where computation actually helps, targets and modalities, what makes a molecule a drug, molecular representation, cheminformatics toolkits, data sources and their quality problems, splitting and leakage, protein structure and structure prediction, docking, molecular dynamics, free energy methods, QSAR and property prediction, ADMET prediction, generative models, active learning and the DMTA loop, benchmarks, validation that means something, pipeline engineering, compute, reproducibility, GxP and computer system validation, and software as a medical device.
Biology and Chemistry Foundations: atoms and periodicity, bonding and molecular geometry, intermolecular forces, phases and solutions, thermodynamics, kinetics, equilibrium and acid-base chemistry, redox and electrochemistry, organic chemistry and reaction mechanism, analytical methods and spectroscopy, biomolecules, enzymes, bioenergetics and metabolism, membranes and transport, cell architecture, the cell cycle and signalling, and evolution — the chemistry and biology a technical reader actually needs, with the reasoning that transfers.
Biomedical Engineering: physiological signal processing, medical imaging modality physics and DICOM internals, clinical data structures, the technicals of clinical machine learning (prevalence, calibration, shortcut learning, the validation ladder), bioinformatics, structural and systems biology, PK/PD, physiological modelling, biomechanics, biomaterials, tissue engineering, neural interfaces and prosthetics, lab automation and instrumentation, and the biostatistics that determine whether any of it holds up.
Access data from the Human Tumor Atlas Network (HTAN) — query the portal database, download from Synapse and Gen3/CRDC, query metadata in BigQuery, and search publications on PubMed