Skip to content
Get packs

PyMC Bayesian Modeling

Builds, checks and compares Bayesian statistical models with PyMC, with diagnostics and reports.

Source Download skill
Repository
K-Dense-AI/scientific-agent-skills
Stars
55,780
Licence
MIT
Files
10 files

What it does

It teaches the agent how to build Bayesian statistical models in PyMC, a Python library, following a careful workflow: define the question, set priors, run sampling, check convergence and predictive fit, then compare models. It includes ready-made templates for linear and hierarchical (multi-group) regression, plus helper scripts that check sampling diagnostics and compare models with PSIS-LOO. Results are reported with uncertainty rather than single numbers.

Great for

  • Analysts or researchers who need probabilistic estimates with uncertainty from their own measurements
  • Developers building hierarchical or multi-group models in PyMC 6.x
  • Students learning Bayesian workflow who want checks (prior predictive, R-hat, ESS, LOO) done correctly

Skip it if

  • Users who need a finished scientific conclusion: the skill states that short test runs are not evidence of converged estimates, and in-sample checks do not validate a model externally.
  • Non-Python environments or older PyMC/ArviZ versions, since it targets PyMC 6.3.2, PyTensor 3.3.2 and ArviZ 1.3 specifically.
  • Sensitive or regulated data handled without review: the agent will execute local Python code on whatever data you provide.

Try asking

  • “Build a hierarchical regression of patient outcomes by clinic and check whether the model converged”
  • “Compare a linear model against a robust Student-t model using PSIS-LOO and report which predicts better”
  • “Predict an outcome for a new, unseen group with uncertainty that includes between-group variation”
  • “Run a prior predictive check on my measurement model before fitting it”
Step-by-step guide: how to use PyMC Bayesian ModelingInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 10 original files, the MIT licence and an attribution note.

    pymc.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip pymc.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip pymc.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Tell the agent your data's units and the question you want answered so it can choose priors in outcome units and define the predictive target.
  • Ask for the full diagnostics report (R-hat, ESS, divergences, BFMI) and for prior sensitivity checks, not just the point estimates.
  • Run the bundled smoke commands first on synthetic data to confirm the environment works, then scale up draws and chains for a real analysis.

What's inside

  • assets/hierarchical_model_template.py
  • assets/linear_regression_template.py
  • references/distributions.md
  • references/model_patterns.md
  • references/sampling_inference.md
  • references/standard_workflow.md
  • references/workflows.md
  • scripts/model_comparison.py
  • scripts/model_diagnostics.py
  • SKILL.md

Redistributed unmodified from K-Dense-AI/scientific-agent-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.