The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The scripts/demo.py file specifies dependencies on standard, well-known data science libraries including marimo, pymc, arviz, numpy, pandas, matplotlib, and scipy. These are established packages typically installed via official registries.
[DYNAMIC_EXECUTION]: The skill uses pymc, which performs runtime compilation of mathematical models into computational graphs (using PyTensor). This is the intended and standard functionality of the library for Bayesian inference.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external datasets (CSV, Parquet) for regression modeling.
Ingestion points: User-provided data files referenced in SKILL.md and loaded in the project scripts.
Boundary markers: Not explicitly present in the modeling code, as the data is processed mathematically rather than via Natural Language instructions.
Capability inventory: The skill performs only standard file reading, data manipulation, and statistical fitting. No network exfiltration or dangerous system commands are triggered by the data contents.
Sanitization: Standard data loading and validation practices are used via pandas.
[COMMAND_EXECUTION]: The documentation includes instructions to execute a demonstration notebook using the marimo edit --sandbox command. The use of the --sandbox flag follows best practices for isolating interactive code execution environment.

bayesian-regression