pymc-modeling skill
pymc-modeling is a set of five PyMC Labs skills that teach an agent to build Bayesian models, check priors and read diagnostics.
- 95 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skills only instruct the agent, while model code and data are processed on your machine.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/pymc-modelingDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills only instruct the agent, while model code and data are processed on your machine.
This is third-party code. Review the repository files before installing.
What it does
pymc-modeling holds five skills in the Agent Skills format. `pymc-modeling` covers model specification, inference, predictions and the model families in pymc-extras. `prior-elicitation` handles prior choice, predictive checks and shrinkage. `arviz-diagnostics` reads MCMC diagnostics, LOO and model comparison. `pytensor-workflows` helps with symbolic graphs, shapes, gradients and custom Ops, while `pymc-mlflow` tracks experiments in MLflow and persists marketing-mix models. Each skill ships reference files and a licence.
Who it is for. Analysts and researchers who build Bayesian models in PyMC and want agent code free of common mistakes.
Good fit when
- A hierarchical model is built for grouped observations.
- Priors are checked for plausible implied outcomes.
- You judge whether LOO and chain convergence can be trusted.
- A shape or gradient error is traced in a PyTensor graph.
Not a fit when
- Classical regression is enough and Bayesian inference is not needed.
- The project uses a library other than PyMC.
- The environment cannot install the Python model packages.
Example request
Build a hierarchical model for the schools data, check priors with predictive checks and show diagnostics.Limitations
The README states that published-repository install commands work only once the manifest is published, and local validation does not prove remote installation. Guidance targets a specific checked version stack and ArviZ DataTree. The skills give instructions and do not run computations for you. The skills cite specific APIs: when PyMC, PyTensor or ArviZ versions change, the references will go stale and need updating. MLflow experiments and marketing-mix persistence require MLflow and PyMC-Marketing to be installed.
How to disable. Delete the skill folders from the agent skills directory or run `/plugin uninstall pymc-modeling@pymc-modeling`. Remove models and MLflow artifacts separately.
FAQ
Plugin or Skills CLI?
Pick one: the description warns that installing both duplicates the skills.
Is a separate service needed?
No. The skills are folders of instructions and references, with no custom installer or running service.
Can I take a single skill?
Yes. Copy the folder you need from `skills/` whole, including references and the licence.
Which licence applies?
MIT.
What does the pymc-mlflow skill do?
It logs experiments to MLflow, stores DataTree artifacts and saves marketing-mix models.
Discussion
No questions or reviews yet.
Related
SkillsA skill for math modeling contests CUMCM, MCM/ICM and the Electrical Cup: 10 stages, three roles, five gates and a LaTeX paper
Eight skills for math modeling contests that carry a task from the problem to a paper with figures and checked references
Probabl skills for machine learning experiments: a scikit-learn and skrub pipeline, evaluation, audit and a link to Skore Hub
A skill that picks a scientific chart from nearly 100 templates for CSV or Excel data and outputs PNG, SVG, PDF and Python code.
A three-stage workflow for math modeling tasks: analysis and model choice, Python or MATLAB computation, and paper assembly in Word or LaTeX
Plugin with 130-plus skills for atomistic materials research: relaxation, dynamics, DFT and structure databases through MCP.
README in short
The README lists the five skills, three install routes and sample prompts, and notes the checked set of library versions.