choosing-a-forecaster
Guides selection of the appropriate skforecast forecaster based on the user's data characteristics and requirements. Provides a decision matrix mapping use cases to forecaster classes. Use when the user is unsure which forecaster to use or asks for a recommendation.
complete-api-reference
Complete constructor signatures and method signatures for all skforecast forecasters, backtesting functions, search functions, cross-validation classes, preprocessing, feature selection, and drift detection. Use when the user needs exact parameter names, types, or defaults for any skforecast class or function.
deep-learning-forecasting
Forecasts time series using recurrent neural networks (RNN, LSTM, GRU) with ForecasterRnn and the create_and_compile_model helper. Covers model architecture, training, and multi-series deep learning. Use when the user wants to use deep learning / neural networks for time series forecasting.
feature-engineering
Creates features for time series forecasting: calendar features with feature_engine (DatetimeFeatures, CyclicalFeatures), rolling statistics with RollingFeatures, differencing, sunlight features, and data scaling. Use when the user wants to improve model accuracy through feature engineering or asks about exogenous variable creation.
forecasting-multiple-series
Forecasts multiple time series simultaneously using a global model with ForecasterRecursiveMultiSeries or ForecasterDirectMultiVariate. Covers data formats, encoding, per-series transformers, and multi-series backtesting. Use when the user has two or more related time series.
forecasting-single-series
Forecasts a single time series using ForecasterRecursive or ForecasterDirect. Covers data preparation, model creation, training, prediction, backtesting, and prediction intervals. Use when the user needs to predict future values of one time series.
hyperparameter-optimization
Optimizes forecaster hyperparameters using grid search, random search, or Bayesian search (Optuna). Covers single-series and multi-series search, cross-validation configuration, and search space definition. Use when the user wants to find the best model configuration.
e2e-auto
Run a fixed OTA regression flow for `examples/v0.81.0` with `agent-device`. Use when the caller wants a built-in end-to-end scenario instead of defining one manually: deploy a known-good OTA bundle, verify a visible UI change plus the deployed `bundleId`, then deploy an intentionally crashing OTA bundle and verify rollback to the previous stable bundle with `RECOVERED` and `crashedBundleId` evidence.
confusion-matrix-generator
Confusion Matrix Generator - Auto-activating skill for ML Training. Triggers on: confusion matrix generator, confusion matrix generator Part of the ML Training skill category.
data-normalization-tool
Data Normalization Tool - Auto-activating skill for ML Training. Triggers on: data normalization tool, data normalization tool Part of the ML Training skill category.
esm
Comprehensive toolkit for protein language models including ESM3 (generative multimodal protein design across sequence, structure, and function) and ESM C (efficient protein embeddings and representations). Use this skill when working with protein sequences, structures, or function prediction; designing novel proteins; generating protein embeddings; performing inverse folding; or conducting protein engineering tasks. Supports both local model usage and cloud-based Forge API for scalable inference.
feature-importance-analyzer
Feature Importance Analyzer - Auto-activating skill for ML Training. Triggers on: feature importance analyzer, feature importance analyzer Part of the ML Training skill category.
lqf-machine-learning-expert-guide
LQF Machine Learning Expert Guide - Auto-activating skill for ML/Statistical Modeling with Critical Discussion Mode. Triggers on: machine learning, modeling, prediction, training, classification, regression, clustering, deep learning, neural network, model evaluation, feature engineering, hyperparameter tuning, overfitting, underfitting, baseline, ablation study, critique my approach, review my model, is this a good idea, should I use, what's wrong with, evaluate my solution, challenge my assumptions, discuss my approach Engages in critical discussion with minimum 3 rounds of iterative refinement. Challenges both user proposals and own suggestions with fact-based critique. Demands evidence and baselines before accepting solutions.
ml-data-leakage-guard
Detects and prevents data leakage in machine learning and mathematical modeling. Auto-activates after ML tasks involving: data cleaning, feature engineering, data augmentation, algorithm development, normalization, missing value imputation, dimensionality reduction, feature selection, time series modeling. Checks if features/statistics would be available at prediction time.
pymc-bayesian-modeling
Bayesian modeling with PyMC. Build hierarchical models, MCMC (NUTS), variational inference, LOO/WAIC comparison, posterior checks, for probabilistic programming and inference.
ralph-loop
Codex-compatible Ralph loop runner with dual engines (compat local state loop + optional open-ralph-wiggum backend).
timesfm-forecasting
Zero-shot time series forecasting with Google's TimesFM foundation model. Use this skill when forecasting ANY univariate time series — sales, sensor readings, stock prices, energy demand, patient vitals, weather, or scientific measurements — without training a custom model. Automatically checks system RAM/GPU before loading the model, supports CSV/DataFrame/array inputs, and returns point forecasts with calibrated prediction intervals. Includes a preflight system checker script that MUST be run before first use to verify the machine can load the model. For classical statistical time series models (ARIMA, SARIMAX, VAR) use statsmodels; for time series classification/clustering use aeon.
ctf-ai-ml
Provides AI and machine learning techniques for CTF challenges. Use when attacking ML models, crafting adversarial examples, performing model extraction, prompt injection, membership inference, training data poisoning, fine-tuning manipulation, neural network analysis, LoRA adapter exploitation, LLM jailbreaking, or solving AI-related puzzles.
devtu-optimize-skills
Optimize ToolUniverse skills for better report quality, evidence handling, and user experience. Apply patterns like tool verification, foundation data layers, disambiguation-first, evidence grading, quantified completeness, and report-only output. Use when reviewing skills, improving existing skills, or creating new ToolUniverse research skills.
tooluniverse-acmg-variant-classification
Systematic ACMG/AMP variant classification using ToolUniverse tools. Given a genetic variant (HGVS, rsID, or gene+change), applies all 28 ACMG criteria (PVS1, PS1-4, PM1-6, PP1-5, BA1, BS1-4, BP1-7) through automated database queries and computational predictions. Produces a final 5-tier classification (Pathogenic / Likely Pathogenic / VUS / Likely Benign / Benign) with evidence summary. Use when asked to classify a variant, interpret a VUS, apply ACMG criteria, assess pathogenicity, or determine clinical significance of a germline variant.
indicator-series
Implement Series-style batch indicators with mathematical precision. Use for new StaticSeries implementations or optimization. Series results are the canonical reference—all other styles must match exactly. Focus on cross-cutting requirements and performance optimization decisions.
continuous-learning
Auto-extract patterns from coding sessions, track corrections, and build reusable knowledge with confidence scoring
