Is it useful for those coming from JAGS or BUGS?

Absolutely. It includes a specific comparison table highlighting key differences in distribution parameterization and execution logic between Stan and BUGS/JAGS.

What version of Stan does this skill support?

This skill is specifically tailored for Stan 2.37 and later, incorporating modern syntax for arrays, complex types, and the latest distribution parameterizations.

Does this skill include R integration examples?

It includes templates for the cmdstanr workflow, including prior/posterior predictive checks and model comparison using the loo package.

Can this skill help resolve Hamiltonian Monte Carlo (HMC) divergences?

Yes, it provides specific patterns for non-centered parameterization and diagnostic checklists to help identify and fix geometry-related sampling issues.

How does this skill handle vectorization?

It provides 'Good vs Bad' code patterns to ensure users implement vectorized probability statements, which significantly improves sampling speed in Stan.

Stan Fundamentals

Name: Stan Fundamentals
Author: choxos

bychoxos

Data Science & ML

Provides foundational knowledge for writing, reviewing, and optimizing high-performance Stan 2.37 Bayesian models.

About

The Stan Fundamentals skill equips Claude with the domain-specific syntax, program architecture, and statistical best practices required to build robust Bayesian models. It covers the essential seven-block program structure, the constrained type system, and critical optimization techniques like vectorization and non-centered parameterization. Whether you are translating models from BUGS/JAGS or building complex hierarchical models from scratch, this skill ensures your Stan code follows modern conventions and integrates seamlessly with workflows like cmdstanr and Statistical Rethinking.

Key Features

Diagnostic checklists for MCMC convergence, Rhat, and Effective Sample Size
Comprehensive reference for constrained types and modern array syntax
Implementation templates for priors, distributions, and generated quantities
Standardized 7-block Stan program structure guidance
0 GitHub stars
Advanced optimization patterns including vectorization and non-centered parameterization

Use Cases

Optimizing slow-running Stan models to improve sampling efficiency and reduce divergences
Translating legacy probabilistic models from BUGS or JAGS into modern Stan syntax
Building new Bayesian hierarchical models from scratch with proper constraints

About

Key Features

Diagnostic checklists for MCMC convergence, Rhat, and Effective Sample Size
Comprehensive reference for constrained types and modern array syntax
Implementation templates for priors, distributions, and generated quantities
Standardized 7-block Stan program structure guidance
0 GitHub stars
Advanced optimization patterns including vectorization and non-centered parameterization

Use Cases

Optimizing slow-running Stan models to improve sampling efficiency and reduce divergences
Translating legacy probabilistic models from BUGS or JAGS into modern Stan syntax
Building new Bayesian hierarchical models from scratch with proper constraints