Cross-reference chemical compound identifiers across 50+ databases (ChEMBL, DrugBank, PubChem, ChEBI, PDB, KEGG) using the UniChem REST API. Resolve InChIKeys to database-specific IDs, find all sources for a compound, discover related compounds by structural connectivity, and batch-translate compound lists between naming systems. No authentication required.
Deep learning framework for drug discovery and materials science. 60+ models (GCN, GAT, AttentiveFP, MPNN, DMPNN, ChemBERTa, GROVER), 50+ molecular featurizers, MoleculeNet benchmarks, hyperparameter optimization, transfer learning. Unified load-featurize-split-train-evaluate API. For fingerprint-only cheminformatics use rdkit-cheminformatics; for featurization hub without training use molfeat-molecular-featurization.
DiffDock is a diffusion-based molecular docking tool that predicts protein-ligand binding poses without requiring a predefined binding site. Use it when the binding site is unknown, when traditional docking fails, or when exploring multiple binding modes. Pipeline: prepare protein (PDB) and ligand (SMILES/SDF) inputs → run DiffDock inference → analyze confidence-ranked poses → visualize in PyMOL or NGLview.
Whole slide image processing for digital pathology. Tissue detection, tile extraction (random, grid, score-based), filter pipelines for H&E/IHC preprocessing. Use for dataset preparation, tile-based deep learning, and slide quality assessment. For advanced spatial proteomics or multiplexed imaging use pathml.
Mass spectrometry spectral matching and metabolite identification with matchms. Use for importing spectra (mzML, MGF, MSP, JSON), filtering/normalizing peaks, computing spectral similarity (cosine, modified cosine, fingerprint), building reproducible processing pipelines, and identifying unknown metabolites from spectral libraries. For full LC-MS/MS proteomics pipelines, use pyopenms instead.
Molecular docking with AutoDock Vina via Python API. Receptor/ligand preparation (Meeko + RDKit), grid box setup, docking execution, pose extraction, binding energy analysis, and batch virtual screening. Use for protein-ligand docking and hit identification.
Query ChEMBL bioactive molecules and drug discovery data using the Python SDK. Search compounds by structure/properties, retrieve bioactivity data (IC50, Ki, EC50), find inhibitors for targets, perform SAR studies, and access drug mechanism/indication data for medicinal chemistry research.
Query ZINC15/ZINC22 virtual compound libraries for drug discovery. Search purchasable lead-like, fragment-like, and drug-like compounds by molecular weight, logP, reactivity class, or SMILES similarity. Download 3D compound sets for docking, retrieve SMILES for in-silico screening. ZINC20 contains 1.4B compounds; purchasable subset is 750M. For bioactivity data use chembl-database-bioactivity; for approved drugs use drugbank-database-access.
Query drug-drug interaction (DDI) data from DDInter via REST API. Search interactions by drug name or ID, retrieve severity levels (major/moderate/minor), interaction mechanisms, and clinical recommendations for drug pairs. Covers 1.7M+ interactions across 2,400+ drugs. No authentication required. For FDA drug labeling use dailymed-database; for pharmacogenomics use clinpgx-database.
Expert patterns for combat systems including hitbox/hurtbox architecture, damage calculation (DamageData class), health components, combat state machines, combo systems, ability cooldowns, and damage popups. Use for action games, RPGs, or fighting games. Trigger keywords: Hitbox, Hurtbox, DamageData, HealthComponent, combat_state, combo_system, ability_cooldown, invincibility_frames, damage_popup.
Expert blueprint for physics queries using RayCast, ShapeCast, and DirectSpaceState. Covers hit detection, volume overlap, mouse picking, and high-performance server-side intersection queries. Use when implementing projectiles, LOS, terrain grounding, or AI sensors. Keywords raycast, shapecast, direct_space_state, intersect_ray, intersect_shape, PhysicsRayQueryParameters, collision mask, mouse picking.
Guide for completing Coq proofs involving arithmetic properties like addition commutativity. This skill should be used when working on Coq proof files that require proving properties about natural number arithmetic using induction, particularly when lemmas like plus_n_O and plus_n_Sm are involved.
Build comprehensive attack trees to visualize threat paths. Use when mapping attack scenarios, identifying defense gaps, or communicating security risks to stakeholders.
Extract patterns and generalizations from multiple observations. Use when detecting recurring themes, building predictive rules, or identifying systemic behaviors from accumulated data. Produces validated patterns with confidence bounds and exception handling.
Implement ReasoningBank adaptive learning with AgentDB for trajectory tracking, verdict judgment, memory distillation, and pattern recognition to build self-learning agents that improve decision-making through experience.
Find substitute materials using CWICR data. Identify equivalent alternatives based on function, cost, and availability.
Screen incoming NDAs and classify them as GREEN (standard), YELLOW (needs review), or RED (significant issues). Use when a new NDA comes in from sales or business development, when assessing NDA risk level, or when deciding whether an NDA needs full counsel review.
Real-time rope/cable physics using Position-Based Dynamics (PBD), Verlet integration, and constraint solvers. Expert in quaternion math, Gauss-Seidel/Jacobi solvers, and tangling detection. Activate on 'rope simulation', 'PBD', 'Position-Based Dynamics', 'Verlet', 'constraint solver', 'quaternion', 'cable dynamics', 'cloth simulation', 'leash physics'. NOT for fluid dynamics (SPH/MPM), fracture simulation (FEM), offline cinematic physics, molecular dynamics, or general game physics engines (use Unity/Unreal built-ins).
Semantic image-text matching with CLIP and alternatives. Use for image search, zero-shot classification, similarity matching. NOT for counting objects, fine-grained classification (celebrities, car models), spatial reasoning, or compositional queries. Activate on "CLIP", "embeddings", "image similarity", "semantic search", "zero-shot classification", "image-text matching".
Analyzes a KiCad PCB to identify high-speed nets by looking up component datasheets via AI. Classifies nets by speed tier (ultra-high/high/medium/low), estimates max frequencies and rise times per interface, and recommends GND return via distances for signal integrity.
Analyzes KiCad PCB files to identify power nets by looking up component datasheets via AI. Use when you need to determine which nets are power/ground nets and what track widths to use, especially when KiCad pintype annotations are missing or unreliable.
Analyzes a KiCad PCB file and creates a comprehensive routing plan. Examines components for fanout needs (BGA/QFN/QFP/PGA), identifies differential pairs, categorizes power/ground nets, and presents a step-by-step routing workflow with explanations.
Comprehensive testing playbook for Local Brain Search memory improvements (Phases 1, 3, 4)
